v4.5 (#403)

2025-07-21 11:43:56 +00:00 · 2023-10-17 10:00:32 +08:00
parent dd8f2744bf
commit 3b776b6639
98 changed files with 1525 additions and 983 deletions
--- a/8
+++ b/8
@@ -9,14 +9,12 @@ ARG name
 # copy packages and one project
 COPY package.json pnpm-lock.yaml pnpm-workspace.yaml ./
 COPY ./packages ./packages
-COPY ./projects/$name ./projects/$name
+COPY ./projects/$name/package.json ./projects/$name/package.json

 RUN \
  [ -f pnpm-lock.yaml ] && pnpm install || \
  (echo "Lockfile not found." && exit 1)

-RUN pnpm prune
-
 # Rebuild the source code only when needed
 FROM node:current-alpine AS builder
 WORKDIR /app
@@ -24,9 +22,11 @@ WORKDIR /app
 ARG name

 # copy common node_modules and one project node_modules
+COPY package.json pnpm-workspace.yaml ./
 COPY --from=deps /app/node_modules ./node_modules
 COPY --from=deps /app/packages ./packages
-COPY --from=deps /app/projects/$name ./projects/$name
+COPY ./projects/$name ./projects/$name
+COPY --from=deps /app/projects/$name/node_modules ./projects/$name/node_modules

 # Uncomment the following line in case you want to disable telemetry during the build.
 ENV NEXT_TELEMETRY_DISABLED 1
--- a/docSite/assets/imgs/v45-1.png
+++ b/docSite/assets/imgs/v45-1.png
--- a/docSite/assets/imgs/v45-2.png
+++ b/docSite/assets/imgs/v45-2.png
--- a/docSite/assets/imgs/v45-3.png
+++ b/docSite/assets/imgs/v45-3.png
--- a/docSite/assets/imgs/v45-4.png
+++ b/docSite/assets/imgs/v45-4.png
--- a/docSite/content/docs/custom-models/chatglm2-m3e.md
+++ b/docSite/content/docs/custom-models/chatglm2-m3e.md
@@ -63,15 +63,15 @@ Authorization 为 sk-aaabbbcccdddeeefffggghhhiiijjjkkk。model 为刚刚在 One

 ```json
 "ChatModels": [
-  //已有模型
+  //其他对话模型
  {
    "model": "chatglm2",
    "name": "chatglm2",
-    "contextMaxToken": 8000,
+    "maxToken": 8000,
+    "price": 0,
    "quoteMaxToken": 4000,
    "maxTemperature": 1.2,
-    "price": 0,
-    "defaultSystem": ""
+    "defaultSystemChatPrompt": ""
  }
 ],
 "VectorModels": [
--- a/docSite/content/docs/custom-models/chatglm2.md
+++ b/docSite/content/docs/custom-models/chatglm2.md
@@ -107,11 +107,11 @@ Authorization 为 sk-aaabbbcccdddeeefffggghhhiiijjjkkk。model 为刚刚在 One
  {
    "model": "chatglm2",
    "name": "chatglm2",
-    "contextMaxToken": 8000,
+    "maxToken": 8000,
+    "price": 0,
    "quoteMaxToken": 4000,
    "maxTemperature": 1.2,
-    "price": 0,
-    "defaultSystem": ""
+    "defaultSystemChatPrompt": ""
  }
 ]
 ```
--- a/docSite/content/docs/development/configuration.md
+++ b/docSite/content/docs/development/configuration.md
@@ -27,31 +27,75 @@ weight: 520
  },
  "ChatModels": [
    {
-      "model": "gpt-3.5-turbo",
-      "name": "GPT35-4k",
-      "contextMaxToken": 4000, // 最大token，均按 gpt35 计算
+      "model": "gpt-3.5-turbo", // 实际调用的模型
+      "name": "GPT35-4k", // 展示的名字
+      "maxToken": 4000, // 最大token，均按 gpt35 计算
      "quoteMaxToken": 2000, // 引用内容最大 token
      "maxTemperature": 1.2, // 最大温度
      "price": 0,
-      "defaultSystem": ""
+      "defaultSystemChatPrompt": ""
    },
    {
      "model": "gpt-3.5-turbo-16k",
      "name": "GPT35-16k",
-      "contextMaxToken": 16000,
+      "maxToken": 16000,
      "quoteMaxToken": 8000,
      "maxTemperature": 1.2,
      "price": 0,
-      "defaultSystem": ""
+      "defaultSystemChatPrompt": ""
    },
    {
      "model": "gpt-4",
      "name": "GPT4-8k",
-      "contextMaxToken": 8000,
+      "maxToken": 8000,
      "quoteMaxToken": 4000,
      "maxTemperature": 1.2,
      "price": 0,
-      "defaultSystem": ""
+      "defaultSystemChatPrompt": ""
+    }
+  ],
+  "QAModel": [ // QA 拆分模型
+    { 
+      "model": "gpt-3.5-turbo-16k",
+      "name": "GPT35-16k",
+      "maxToken": 16000,
+      "price": 0
+    }
+  ],
+  "ExtractModels": [ // 内容提取模型
+    { 
+      "model": "gpt-3.5-turbo-16k",
+      "name": "GPT35-16k",
+      "maxToken": 16000, 
+      "price": 0,
+      "functionCall": true, // 是否支持 function call
+      "functionPrompt": "" // 自定义非 function call 提示词
+    }
+  ],
+  "CQModels": [ // Classify Question: 问题分类模型
+    {
+      "model": "gpt-3.5-turbo-16k",
+      "name": "GPT35-16k",
+      "maxToken": 16000,
+      "price": 0,
+      "functionCall": true,
+      "functionPrompt": ""
+    },
+    {
+      "model": "gpt-4",
+      "name": "GPT4-8k",
+      "maxToken": 8000,
+      "price": 0,
+      "functionCall": true,
+      "functionPrompt": ""
+    }
+  ],
+  "QGModels": [ // Question Generation: 生成下一步指引模型
+    { 
+      "model": "gpt-3.5-turbo",
+      "name": "GPT35-4k",
+      "maxToken": 4000,
+      "price": 0
    }
  ],
  "VectorModels": [
@@ -62,36 +106,6 @@ weight: 520
      "defaultToken": 500,
      "maxToken": 3000
    }
-  ],
-  "QAModel": { // QA 拆分模型
-    "model": "gpt-3.5-turbo-16k",
-    "name": "GPT35-16k",
-    "maxToken": 16000,
-    "price": 0
-  },
-  "ExtractModel": { // 内容提取模型
-    "model": "gpt-3.5-turbo-16k",
-    "functionCall": true, // 是否使用 functionCall
-    "name": "GPT35-16k",
-    "maxToken": 16000,
-    "price": 0,
-    "prompt": ""
-  },
-  "CQModel": { // Classify Question: 问题分类模型
-    "model": "gpt-3.5-turbo-16k",
-    "functionCall": true,
-    "name": "GPT35-16k",
-    "maxToken": 16000,
-    "price": 0,
-    "prompt": ""
-  },
-  "QGModel": { // Question Generation: 生成下一步指引模型
-    "model": "gpt-3.5-turbo",
-    "name": "GPT35-4k",
-    "maxToken": 4000,
-    "price": 0,
-    "prompt": "",
-    "functionCall": false
-  }
+  ]
 }
 ```
--- a/docSite/content/docs/installation/docker.md
+++ b/docSite/content/docs/installation/docker.md
@@ -139,6 +139,21 @@ docker-compose 端口定义为：`映射端口:运行端口`。

 （自行补习 docker 基本知识）

+### relation "modeldata" does not exist
+
+PG 数据库没有连接上/初始化失败，可以查看日志。FastGPT 会在每次连接上 PG 时进行表初始化，如果报错会有对应日志。
+
+1. 检查数据库容器是否正常启动
+2. 非 docker 部署的，需要手动安装 pg vector 插件
+3. 查看 fastgpt 日志，有没有相关报错
+
+### Operation `auth_codes.findOne()` buffering timed out after 10000ms
+
+mongo连接失败，检查
+1. mongo 服务有没有起来(有些 cpu 不支持 AVX，无法用 mongo5，需要换成 mongo4.x，可以dockerhub找个最新的4.x，修改镜像版本，重新运行）
+2. 环境变量（账号密码，注意host和port）
+
+
 ### 错误排查方式

 遇到问题先按下面方式排查。
--- a/docSite/content/docs/installation/one-api.md
+++ b/docSite/content/docs/installation/one-api.md
@@ -99,12 +99,12 @@ CHAT_API_KEY=sk-xxxxxx
    {
      "model": "ERNIE-Bot", // 这里的模型需要对应 One API 的模型
      "name": "文心一言", // 对外展示的名称
-      "contextMaxToken": 4000, // 最大长下文 token，无论什么模型都按 GPT35 的计算。GPT 外的模型需要自行大致计算下这个值。可以调用官方接口去比对 Token 的倍率，然后在这里粗略计算。
+      "maxToken": 4000, // 最大长下文 token，无论什么模型都按 GPT35 的计算。GPT 外的模型需要自行大致计算下这个值。可以调用官方接口去比对 Token 的倍率，然后在这里粗略计算。
      // 例如：文心一言的中英文 token 基本是 1:1，而 GPT 的中文 Token 是 2:1，如果文心一言官方最大 Token 是 4000，那么这里就可以填 8000，保险点就填 7000.
+      "price": 0, // 1个token 价格 => 1.5 / 100000 * 1000 = 0.015元/1k token
      "quoteMaxToken": 2000, // 引用知识库的最大 Token
      "maxTemperature": 1, // 最大温度
-      "price": 0, // 1个token 价格 => 1.5 / 100000 * 1000 = 0.015元/1k token
-      "defaultSystem": "" // 默认的系统提示词
+      "defaultSystemChatPrompt": "" // 默认的系统提示词
    }
    ...
 ],
--- a/docSite/content/docs/installation/upgrading/45.md
+++ b/docSite/content/docs/installation/upgrading/45.md
@@ -0,0 +1,84 @@
+---
+title: 'V4.5(需进行较为复杂更新)'
+description: 'FastGPT V4.5 更新'
+icon: 'upgrade'
+draft: false
+toc: true
+weight: 839
+---
+
+FastGPT V4.5 引入 PgVector0.5 版本的 HNSW 索引，极大的提高了知识库检索的速度，比起`IVFFlat`索引大致有3~10倍的性能提升，可轻松实现百万数据毫秒级搜索。缺点在于构建索引的速度非常慢，4c16g 500w 组数据使用`并行构建`大约花了 48 小时。具体参数配置可参考 [PgVector官方](https://github.com/pgvector/pgvector)
+
+下面需要对数据库进行一些操作升级：
+
+## PgVector升级：Sealos 部署方案
+
+1. 点击[Sealos桌面](https://cloud.sealos.io)的数据库应用。
+2. 点击【pg】数据库的详情。
+3. 点击右上角的重启，等待重启完成。
+4. 点击左侧的一键链接，等待打开 Terminal。
+5. 依次输入下方 sql 命令
+
+```sql
+-- 升级插件名
+ALTER EXTENSION vector UPDATE;
+-- 插件是否升级成功，成功的话，vector插件版本为 0.5.0，旧版的为 0.4.1
+\dx
+
+-- 下面两个语句会设置 pg 在构建索引时可用的内存大小，需根据自身的数据库规格来动态配置，可配置为 1/4 的内存大小
+alter system set maintenance_work_mem = '2400MB'; 
+select pg_reload_conf();
+
+-- 开始构建索引，该索引构建时间非常久，直接点击右上角的叉，退出 Terminal 即可
+CREATE INDEX CONCURRENTLY vector_index ON modeldata USING hnsw (vector vector_ip_ops) WITH (m = 16, ef_construction = 64);
+-- 可以再次点击一键链接，进入 Terminal，输入下方命令，如果看到 "vector_index" hnsw (vector vector_ip_ops) WITH (m='16', ef_construction='64') 则代表构建完成（注意，后面没有 INVALID）
+\d modeldata
+```
+
+|                       |                       |
+| --------------------- | --------------------- |
+| ![](/imgs/v45-1.png) | ![](/imgs/v45-2.png) |
+| ![](/imgs/v45-3.png) | ![](/imgs/v45-4.png) |
+
+
+
+## PgVector升级：Docker-compose.yml 部署方案
+
+下面的命令是基于给的 docker-compose 模板，如果数据库账号密码更换了，请自行调整。
+
+1. 修改 `docker-compose.yml` 中pg的镜像版本，改成 `ankane/pgvector:v0.5.0` 或 `registry.cn-hangzhou.aliyuncs.com/fastgpt/pgvector:v0.5.0`
+2. 重启 pg 容器(docker-compose pull && docker-compose up -d)，等待重启完成。
+3. 进入容器: `docker exec -it pg bash`
+4. 连接数据库: `psql 'postgresql://username:password@localhost:5432/postgres'`
+5. 执行下面 sql 命令
+
+```sql
+-- 升级插件名
+ALTER EXTENSION vector UPDATE;
+-- 插件是否升级成功，成功的话，vector插件版本为 0.5.0，旧版的为 0.4.2
+\dx
+
+-- 下面两个语句会设置 pg 在构建索引时可用的内存大小，需根据自身的数据库规格来动态配置，可配置为 1/4 的内存大小
+alter system set maintenance_work_mem = '2400MB'; 
+select pg_reload_conf();
+
+-- 开始构建索引，该索引构建时间非常久，直接关掉终端即可，不要使用 ctrl+c 关闭
+CREATE INDEX CONCURRENTLY vector_index ON modeldata USING hnsw (vector vector_ip_ops) WITH (m = 16, ef_construction = 64);
+-- 可以再次连接数据库，输入下方命令。如果看到 "vector_index" hnsw (vector vector_ip_ops) WITH (m='16', ef_construction='64') 则代表构建完成（注意，后面没有 INVALID）
+\d modeldata
+```
+
+## 版本新功能介绍
+
+### Fast GPT V4.5
+
+1. 新增 - 升级 PgVector 插件，引入 HNSW 索引，极大加快的知识库搜索速度。
+2. 新增 - AI对话模块，增加【返回AI内容】选项，可控制 AI 的内容不直接返回浏览器。
+3. 新增 - 支持问题分类选择模型
+4. 优化 - TextSplitter，采用递归拆解法。
+5. 优化 - 高级编排 UX 性能
+6. 修复 - 分享链接鉴权问题
+
+## 该版本需要修改 `config.json` 文件
+
+最新配置可参考: [V45版本最新 config.json](/docs/development/configuration)
--- a/docSite/content/docs/use-cases/ai_settings.md
+++ b/docSite/content/docs/use-cases/ai_settings.md
@@ -0,0 +1,94 @@
+---
+title: "AI 高级配置说明"
+description: "FastGPT AI 高级配置说明"
+icon: "sign_language"
+draft: false
+toc: true
+weight: 310
+---
+
+在 FastGPT 的 AI 对话模块中，有一个 AI 高级配置，里面包含了 AI 模型的参数配置，本文详细介绍这些配置的含义。
+
+# 返回AI内容
+
+这是一个开关，打开的时候，当 AI 对话模块运行时，会将其输出的内容返回到浏览器（API响应）；如果关闭，AI 输出的内容不会返回到浏览器，但是生成的内容仍可以通过【AI回复】进行输出。你可以将【AI回复】连接到其他模块中。
+
+# 温度
+
+可选范围0-10，约大代表生成的内容约自由扩散，越小代表约严谨。调节能力有限，知识库问答场景通常设置为0。
+
+# 回复上限
+
+控制 AI 回复的最大 Tokens，较小的值可以一定程度上减少 AI 的废话，但也可能导致 AI 回复不完整。
+
+# 引用模板 & 引用提示词
+
+这两个参数与知识库问答场景相关，可以控制知识库相关的提示词。
+
+## AI 对话消息组成
+
+想使用明白这两个变量，首先要了解传递传递给 AI 模型的消息格式。它是一个数组，FastGPT 中这个数组的组成形式为：
+
+```json
+[
+    内置提示词（config.json 配置，一般为空）
+    系统提示词 （用户输入的提示词）
+    历史记录
+    问题（由引用提示词、引用模板和用户问题组成）
+]
+```
+
+{{% alert icon="🍅" context="success" %}}
+Tips: 可以通过点击上下文按键查看完整的
+{{% /alert %}}
+
+## 引用模板和提示词设计
+
+引用模板和引用提示词通常是成对出现，引用提示词依赖引用模板。
+
+FastGPT 知识库采用 QA 对(不一定都是问答格式，仅代表两个变量)的格式存储，在转义成字符串时候会根据**引用模板**来进行格式化。知识库包含 3 个变量： q, a, file_id, index, source，可以通过 {{q}} {{a}} {{file_id}} {{index}} {{source}} 按需引入。下面一个模板例子：
+
+**引用模板**
+
+```
+{instruction:"{{q}}",output:"{{a}}",source:"{{source}}"}
+```
+
+搜索到的知识库，会自动将 q,a,source 替换成对应的内容。每条搜索到的内容，会通过 `\n` 隔开。例如：
+```
+{instruction:"电影《铃芽之旅》的导演是谁？",output:"电影《铃芽之旅》的导演是新海诚。",source:"手动输入"}
+{instruction:"本作的主人公是谁？",output:"本作的主人公是名叫铃芽的少女。",source:""}
+{instruction:"电影《铃芽之旅》男主角是谁？",output:"电影《铃芽之旅》男主角是宗像草太，由松村北斗配音。",source:""}
+{instruction:"电影《铃芽之旅》的编剧是谁？22",output:"新海诚是本片的编剧。",source:"手动输入"}
+```
+
+**引用提示词**
+
+引用模板需要和引用提示词一起使用，提示词中可以写引用模板的格式说明以及对话的要求等。可以使用 {{quote}} 来使用 **引用模板**，使用 {{question}} 来引入问题。例如：
+
+```
+你的背景知识:
+"""
+{{quote}}
+"""
+对话要求：
+1. 背景知识是最新的，其中 instruction 是相关介绍，output 是预期回答或补充。
+2. 使用背景知识回答问题。
+3. 背景知识无法回答问题时，你可以礼貌的的回答用户问题。
+我的问题是:"{{question}}"
+```
+
+转义后则为：
+```
+你的背景知识:
+"""
+{instruction:"电影《铃芽之旅》的导演是谁？",output:"电影《铃芽之旅》的导演是新海诚。",source:"手动输入"}
+{instruction:"本作的主人公是谁？",output:"本作的主人公是名叫铃芽的少女。",source:""}
+{instruction:"电影《铃芽之旅》男主角是谁？",output:"电影《铃芽之旅》男主角是宗像草太，由松村北斗配音}
+"""
+对话要求：
+1. 背景知识是最新的，其中 instruction 是相关介绍，output 是预期回答或补充。
+2. 使用背景知识回答问题。
+3. 背景知识无法回答问题时，你可以礼貌的的回答用户问题。
+我的问题是:"{{question}}"
+```
--- a/docSite/content/docs/use-cases/prompt.md
+++ b/docSite/content/docs/use-cases/prompt.md
@@ -1,109 +0,0 @@
---
-title: "提示词 & 引用提示词"
-description: "FastGPT 提示词 & 引用提示词说明"
-icon: "sign_language"
-draft: false
-toc: true
-weight: 310
---
-
-限定词从 V4.4.3 版本后去除，被“引用提示词”和“引用模板”替代。
-
-# AI 对话消息组成
-
-传递给 AI 模型的消息是一个数组，FastGPT 中这个数组的组成形式为：
-
-```json
-[
-    内置提示词（config.json 配置，一般为空）
-    提示词 （用户输入的提示词）
-    历史记录
-    问题（会由输入的问题、引用提示词和引用模板来决定）
-]
-```
-
-{{% alert icon="🍅" context="success" %}}
-Tips: 可以通过点击上下文按键查看完整的
-{{% /alert %}}
-
-# 引用模板和提示词设计
-
-知识库采用 QA 对的格式存储，在转义成字符串时候会根据**引用模板**来进行格式化。知识库包含 3 个变量： q,a 和 source，可以通过 {{q}} {{a}} {{source}} 按需引入。下面一个模板例子：
-
-**引用模板**
-
-```
-{instruction:"{{q}}",output:"{{a}}",source:"{{source}}"}
-```
-
-搜索到的知识库，会自动将 q,a,source 替换成对应的内容。每条搜索到的内容，会通过 `\n` 隔开。例如：
-```
-{instruction:"电影《铃芽之旅》的导演是谁？",output:"电影《铃芽之旅》的导演是新海诚。",source:"手动输入"}
-{instruction:"本作的主人公是谁？",output:"本作的主人公是名叫铃芽的少女。",source:""}
-{instruction:"电影《铃芽之旅》男主角是谁？",output:"电影《铃芽之旅》男主角是宗像草太，由松村北斗配音。",source:""}
-{instruction:"电影《铃芽之旅》的编剧是谁？22",output:"新海诚是本片的编剧。",source:"手动输入"}
-```
-
-**引用提示词**
-
-引用模板需要和引用提示词一起使用，提示词中可以写引用模板的格式说明以及对话的要求等。可以使用 {{quote}} 来使用 **引用模板**，使用 {{question}} 来引入问题。例如：
-
-```
-你的背景知识:
-"""
-{{quote}}
-"""
-对话要求：
-1. 背景知识是最新的，其中 instruction 是相关介绍，output 是预期回答或补充。
-2. 使用背景知识回答问题。
-3. 背景知识无法回答问题时，你可以礼貌的的回答用户问题。
-我的问题是:"{{question}}"
-```
-
-
-# 提示词案例
-
-## 仅回复知识库里的内容
-
-**引用提示词**里添加：
-```
-你的背景知识:
-"""
-{{quote}}
-"""
-对话要求：
-1. 回答前，请先判断背景知识是否足够回答问题，如果无法回答，请直接回复：“对不起，我无法回答你的问题~”。
-2. 背景知识是最新的，其中 instruction 是相关介绍，output 是预期回答或补充。
-3. 使用背景知识回答问题。
-我的问题是:"{{question}}"
-```
-
-## 说明引用来源
-
-**引用模板:**
-
-```
-{instruction:"{{q}}",output:"{{a}}",source:"{{source}}"}
-```
-
-**引用提示词：**
-
-```
-你的背景知识:
-"""
-{{quote}}
-"""
-对话要求：
-1. 背景知识是最新的，其中 instruction 是相关介绍，output 是预期回答或补充，source是背景来源。
-2. 使用背景知识回答问题。
-3. 在回答问题后，你需要给出本次回答对应的背景来源，来源展示格式如下：
-
-“
-这是AI作答。本次知识来源：
-1. source1
-2. source2
-......
-”
-
-我的问题是:"{{question}}"
-```
--- a/docSite/content/docs/workflow/examples/fixingEvidence.md
+++ b/docSite/content/docs/workflow/examples/fixingEvidence.md
@@ -232,7 +232,7 @@ weight: 142
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "AI回复",
        "description": "将在 stream 回复完毕后触发",
        "valueType": "string",
        "type": "source",
--- a/docSite/content/docs/workflow/examples/google_search.md
+++ b/docSite/content/docs/workflow/examples/google_search.md
@@ -432,7 +432,7 @@ export default async function (ctx: FunctionContext) {
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "AI回复",
        "description": "直接响应，无需配置",
        "type": "hidden",
        "targets": []
--- a/docSite/content/docs/workflow/examples/lab_appointment.md
+++ b/docSite/content/docs/workflow/examples/lab_appointment.md
@@ -751,7 +751,7 @@ HTTP 模块允许你调用任意 POST 类型的 HTTP 接口，从而实验一些
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "模型AI回复回复",
        "description": "将在 stream 回复完毕后触发",
        "valueType": "string",
        "type": "source",
--- a/docSite/content/docs/workflow/examples/op_question.md
+++ b/docSite/content/docs/workflow/examples/op_question.md
@@ -313,7 +313,7 @@ weight: 144
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "AI回复",
        "description": "将在 stream 回复完毕后触发",
        "valueType": "string",
        "type": "source",
--- a/docSite/content/docs/workflow/examples/versatile_assistant.md
+++ b/docSite/content/docs/workflow/examples/versatile_assistant.md
@@ -745,7 +745,7 @@ PS2：配置中的问题分类还包含着“联网搜索”，这个是另一
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "AI回复",
        "description": "将在 stream 回复完毕后触发",
        "valueType": "string",
        "type": "source",
@@ -903,7 +903,7 @@ PS2：配置中的问题分类还包含着“联网搜索”，这个是另一
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "AI回复",
        "description": "将在 stream 回复完毕后触发",
        "valueType": "string",
        "type": "source",
@@ -1117,7 +1117,7 @@ PS2：配置中的问题分类还包含着“联网搜索”，这个是另一
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "AI回复",
        "description": "将在 stream 回复完毕后触发",
        "valueType": "string",
        "type": "source",
@@ -1484,7 +1484,7 @@ PS2：配置中的问题分类还包含着“联网搜索”，这个是另一
    "outputs": [
      {
        "key": "answerText",
-        "label": "模型回复",
+        "label": "AI回复",
        "description": "将在 stream 回复完毕后触发",
        "valueType": "string",
        "type": "source",
--- a/packages/common/mongo/init.ts
+++ b/packages/common/mongo/init.ts
@@ -29,7 +29,9 @@ export async function connectMongo({
      bufferCommands: true,
      maxConnecting: Number(process.env.DB_MAX_LINK || 5),
      maxPoolSize: Number(process.env.DB_MAX_LINK || 5),
-      minPoolSize: 2
+      minPoolSize: 2,
+      connectTimeoutMS: 20000,
+      waitQueueTimeoutMS: 20000
    });

    console.log('mongo connected');
--- a/packages/common/package.json
+++ b/packages/common/package.json
@@ -5,7 +5,9 @@
    "mongoose": "^7.0.2",
    "winston": "^3.10.0",
    "winston-mongodb": "^5.1.1",
-    "axios": "^1.5.1"
+    "axios": "^1.5.1",
+    "nextjs-cors": "^2.1.2",
+    "next": "13.5.2"
  },
  "devDependencies": {
    "@types/node": "^20.8.5"
--- a/packages/common/tools/nextjs.ts
+++ b/packages/common/tools/nextjs.ts
@@ -0,0 +1,19 @@
+import type { NextApiResponse, NextApiHandler, NextApiRequest } from 'next';
+import NextCors from 'nextjs-cors';
+
+export function withNextCors(handler: NextApiHandler): NextApiHandler {
+  return async function nextApiHandlerWrappedWithNextCors(
+    req: NextApiRequest,
+    res: NextApiResponse
+  ) {
+    const methods = ['GET', 'eHEAD', 'PUT', 'PATCH', 'POST', 'DELETE'];
+    const origin = req.headers.origin;
+    await NextCors(req, res, {
+      methods,
+      origin: origin,
+      optionsSuccessStatus: 200
+    });
+
+    return handler(req, res);
+  };
+}
--- a/packages/common/tools/str.ts
+++ b/packages/common/tools/str.ts
@@ -13,20 +13,10 @@ export const hashStr = (psw: string) => {
 /* simple text, remove chinese space and extra \n */
 export const simpleText = (text: string) => {
  text = text.replace(/([\u4e00-\u9fa5])[\s&&[^\n]]+([\u4e00-\u9fa5])/g, '$1$2');
-  text = text.replace(/\n{2,}/g, '\n');
+  text = text.replace(/\n{3,}/g, '\n\n');
  text = text.replace(/[\s&&[^\n]]{2,}/g, ' ');
  text = text.replace(/[\x00-\x08]/g, ' ');
+  text = text.replace(/\r\n|\r/g, '\n');

-  // replace empty \n
-  let newText = '';
-  let lastChar = '';
-  for (let i = 0; i < text.length; i++) {
-    const currentChar = text[i];
-    if (currentChar === '\n' && !/[。？！；.?!;]/g.test(lastChar)) {
-    } else {
-      newText += currentChar;
-    }
-    lastChar = currentChar;
-  }
-  return newText;
+  return text;
 };
--- a/projects/app/src/service/common/stream.ts
+++ b/projects/app/src/service/common/stream.ts
--- a/packages/core/ai/config.ts
+++ b/packages/core/ai/config.ts
@@ -11,6 +11,7 @@ export const getAIApi = (props?: UserModelSchema['openaiAccount'], timeout = 600
    apiKey: props?.key || systemAIChatKey,
    baseURL: props?.baseUrl || baseUrl,
    httpAgent: global.httpsAgent,
-    timeout
+    timeout,
+    maxRetries: 2
  });
 };
--- a/packages/core/ai/type.d.ts
+++ b/packages/core/ai/type.d.ts
@@ -4,3 +4,9 @@ export type ChatCompletion = OpenAI.Chat.ChatCompletion;
 export type CreateChatCompletionRequest = OpenAI.Chat.ChatCompletionCreateParams;

 export type StreamChatType = Stream<OpenAI.Chat.ChatCompletionChunk>;
+
+export type PromptTemplateItem = {
+  title: string;
+  desc: string;
+  value: string;
+};
--- a/packages/core/package.json
+++ b/packages/core/package.json
@@ -5,7 +5,7 @@
    "@fastgpt/common": "workspace:*",
    "@fastgpt/support": "workspace:*",
    "encoding": "^0.1.13",
-    "openai": "^4.11.1",
+    "openai": "^4.12.1",
    "tunnel": "^0.0.6"
  },
  "devDependencies": {
--- a/packages/support/outLink/auth.ts
+++ b/packages/support/outLink/auth.ts
@@ -63,5 +63,6 @@ export type AuthShareChatInitProps = {
 };

 export function authShareChatInit(data: AuthShareChatInitProps) {
+  if (!global.feConfigs?.isPlus) return;
  return POST('/support/outLink/authShareChatInit', data);
 }
--- a/packages/support/package.json
+++ b/packages/support/package.json
@@ -5,7 +5,8 @@
    "@fastgpt/common": "workspace:*",
    "cookie": "^0.5.0",
    "jsonwebtoken": "^9.0.2",
-    "axios": "^1.5.1"
+    "axios": "^1.5.1",
+    "next": "13.5.2"
  },
  "devDependencies": {
    "@types/cookie": "^0.5.2",
--- a/packages/support/user/auth.ts
+++ b/packages/support/user/auth.ts
@@ -1,8 +1,8 @@
+import type { NextApiResponse, NextApiRequest } from 'next';
 import Cookie from 'cookie';
-import { authJWT } from './tools';
+import jwt from 'jsonwebtoken';
 import { authOpenApiKey } from '../openapi/auth';
 import { authOutLinkId } from '../outLink/auth';
-
 import { MongoUser } from './schema';
 import type { UserModelSchema } from './type.d';
 import { ERROR_ENUM } from '@fastgpt/common/constant/errorCode';
@@ -39,7 +39,7 @@ export const authUser = async ({
  authBalance = false,
  authOutLink
 }: {
-  req: any;
+  req: NextApiRequest;
  authToken?: boolean;
  authRoot?: boolean;
  authApiKey?: boolean;
@@ -165,3 +165,42 @@ export const authUser = async ({
    apikey: openApiKey
  };
 };
+
+/* 生成 token */
+export function generateToken(userId: string) {
+  const key = process.env.TOKEN_KEY as string;
+  const token = jwt.sign(
+    {
+      userId,
+      exp: Math.floor(Date.now() / 1000) + 60 * 60 * 24 * 7
+    },
+    key
+  );
+  return token;
+}
+// auth token
+export function authJWT(token: string) {
+  return new Promise<string>((resolve, reject) => {
+    const key = process.env.TOKEN_KEY as string;
+
+    jwt.verify(token, key, function (err, decoded: any) {
+      if (err || !decoded?.userId) {
+        reject(ERROR_ENUM.unAuthorization);
+        return;
+      }
+      resolve(decoded.userId);
+    });
+  });
+}
+
+/* set cookie */
+export const setCookie = (res: NextApiResponse, token: string) => {
+  res.setHeader(
+    'Set-Cookie',
+    `token=${token}; Path=/; HttpOnly; Max-Age=604800; Samesite=None; Secure;`
+  );
+};
+/* clear cookie */
+export const clearCookie = (res: NextApiResponse) => {
+  res.setHeader('Set-Cookie', 'token=; Path=/; Max-Age=0');
+};
--- a/packages/support/user/tools.ts
+++ b/packages/support/user/tools.ts
@@ -1,28 +0,0 @@
-import jwt from 'jsonwebtoken';
-import { ERROR_ENUM } from '@fastgpt/common/constant/errorCode';
-
-/* 生成 token */
-export const generateToken = (userId: string) => {
-  const key = process.env.TOKEN_KEY as string;
-  const token = jwt.sign(
-    {
-      userId,
-      exp: Math.floor(Date.now() / 1000) + 60 * 60 * 24 * 7
-    },
-    key
-  );
-  return token;
-};
-// auth token
-export const authJWT = (token: string) =>
-  new Promise<string>((resolve, reject) => {
-    const key = process.env.TOKEN_KEY as string;
-
-    jwt.verify(token, key, function (err, decoded: any) {
-      if (err || !decoded?.userId) {
-        reject(ERROR_ENUM.unAuthorization);
-        return;
-      }
-      resolve(decoded.userId);
-    });
-  });
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -35,6 +35,12 @@ importers:
      mongoose:
        specifier: ^7.0.2
        version: registry.npmmirror.com/mongoose@7.0.2
+      next:
+        specifier: 13.5.2
+        version: registry.npmmirror.com/next@13.5.2(@babel/core@7.23.2)(react-dom@18.2.0)(react@18.2.0)(sass@1.58.3)
+      nextjs-cors:
+        specifier: ^2.1.2
+        version: registry.npmmirror.com/nextjs-cors@2.1.2(next@13.5.2)
      winston:
        specifier: ^3.10.0
        version: registry.npmmirror.com/winston@3.10.0
@@ -58,8 +64,8 @@ importers:
        specifier: ^0.1.13
        version: registry.npmmirror.com/encoding@0.1.13
      openai:
-        specifier: ^4.11.1
-        version: registry.npmmirror.com/openai@4.11.1(encoding@0.1.13)
+        specifier: ^4.12.1
+        version: registry.npmmirror.com/openai@4.12.1(encoding@0.1.13)
      tunnel:
        specifier: ^0.0.6
        version: registry.npmmirror.com/tunnel@0.0.6
@@ -82,6 +88,9 @@ importers:
      jsonwebtoken:
        specifier: ^9.0.2
        version: registry.npmmirror.com/jsonwebtoken@9.0.2
+      next:
+        specifier: 13.5.2
+        version: registry.npmmirror.com/next@13.5.2(@babel/core@7.23.2)(react-dom@18.2.0)(react@18.2.0)(sass@1.58.3)
    devDependencies:
      '@types/cookie':
        specifier: ^0.5.2
@@ -200,9 +209,6 @@ importers:
      next-i18next:
        specifier: ^14.0.0
        version: registry.npmmirror.com/next-i18next@14.0.3(i18next@23.5.1)(next@13.5.2)(react-i18next@13.2.2)(react@18.2.0)
-      nextjs-cors:
-        specifier: ^2.1.2
-        version: registry.npmmirror.com/nextjs-cors@2.1.2(next@13.5.2)
      nprogress:
        specifier: ^0.2.0
        version: registry.npmmirror.com/nprogress@0.2.0
@@ -288,6 +294,9 @@ importers:
      '@types/multer':
        specifier: ^1.4.7
        version: registry.npmmirror.com/@types/multer@1.4.7
+      '@types/node':
+        specifier: ^20.8.5
+        version: registry.npmmirror.com/@types/node@20.8.5
      '@types/papaparse':
        specifier: ^5.3.7
        version: registry.npmmirror.com/@types/papaparse@5.3.7
@@ -9581,11 +9590,11 @@ packages:
      mimic-fn: registry.npmmirror.com/mimic-fn@4.0.0
    dev: true

-  registry.npmmirror.com/openai@4.11.1(encoding@0.1.13):
-    resolution: {integrity: sha512-GU0HQWbejXuVAQlDjxIE8pohqnjptFDIm32aPlNT1H9ucMz1VJJD0DaTJRQsagNaJ97awWjjVLEG7zCM6sm4SA==, registry: https://registry.npm.taobao.org/, tarball: https://registry.npmmirror.com/openai/-/openai-4.11.1.tgz}
-    id: registry.npmmirror.com/openai/4.11.1
+  registry.npmmirror.com/openai@4.12.1(encoding@0.1.13):
+    resolution: {integrity: sha512-EAoUwm4dtiWvFwBhOCK/VfF8sj1ZU8+aAIJnfT4NyeTfrt1DM/6Gdd6fOZWTjBYryTAqu9Vpb5+9Wu6JMtm/gA==, registry: https://registry.npm.taobao.org/, tarball: https://registry.npmmirror.com/openai/-/openai-4.12.1.tgz}
+    id: registry.npmmirror.com/openai/4.12.1
    name: openai
-    version: 4.11.1
+    version: 4.12.1
    hasBin: true
    dependencies:
      '@types/node': registry.npmmirror.com/@types/node@18.18.5
--- a/projects/app/data/config.json
+++ b/projects/app/data/config.json
@@ -8,68 +8,85 @@
    {
      "model": "gpt-3.5-turbo",
      "name": "GPT35-4k",
-      "contextMaxToken": 4000,
+      "price": 0,
+      "maxToken": 4000,
      "quoteMaxToken": 2000,
      "maxTemperature": 1.2,
-      "price": 0,
-      "defaultSystem": ""
+      "censor": false,
+      "defaultSystemChatPrompt": ""
    },
    {
      "model": "gpt-3.5-turbo-16k",
      "name": "GPT35-16k",
-      "contextMaxToken": 16000,
+      "maxToken": 16000,
+      "price": 0,
      "quoteMaxToken": 8000,
      "maxTemperature": 1.2,
-      "price": 0,
-      "defaultSystem": ""
+      "censor": false,
+      "defaultSystemChatPrompt": ""
    },
    {
      "model": "gpt-4",
      "name": "GPT4-8k",
-      "contextMaxToken": 8000,
+      "maxToken": 8000,
+      "price": 0,
      "quoteMaxToken": 4000,
      "maxTemperature": 1.2,
+      "censor": false,
+      "defaultSystemChatPrompt": ""
+    }
+  ],
+  "QAModels": [
+    {
+      "model": "gpt-3.5-turbo-16k",
+      "name": "GPT35-16k",
+      "maxToken": 16000,
+      "price": 0
+    }
+  ],
+  "CQModels": [
+    {
+      "model": "gpt-3.5-turbo-16k",
+      "name": "GPT35-16k",
+      "maxToken": 16000,
      "price": 0,
-      "defaultSystem": ""
+      "functionCall": true,
+      "functionPrompt": ""
+    },
+    {
+      "model": "gpt-4",
+      "name": "GPT4-8k",
+      "maxToken": 8000,
+      "price": 0,
+      "functionCall": true,
+      "functionPrompt": ""
+    }
+  ],
+  "ExtractModels": [
+    {
+      "model": "gpt-3.5-turbo-16k",
+      "name": "GPT35-16k",
+      "maxToken": 16000,
+      "price": 0,
+      "functionCall": true,
+      "functionPrompt": ""
+    }
+  ],
+  "QGModels": [
+    {
+      "model": "gpt-3.5-turbo",
+      "name": "GPT35-4K",
+      "maxToken": 4000,
+      "price": 0
    }
  ],
  "VectorModels": [
    {
      "model": "text-embedding-ada-002",
      "name": "Embedding-2",
-      "price": 0,
-      "defaultToken": 500,
+      "price": 0.2,
+      "defaultToken": 700,
      "maxToken": 3000
    }
-  ],
-  "QAModel": {
-    "model": "gpt-3.5-turbo-16k",
-    "name": "GPT35-16k",
-    "maxToken": 16000,
-    "price": 0
-  },
-  "ExtractModel": {
-    "model": "gpt-3.5-turbo-16k",
-    "functionCall": true,
-    "name": "GPT35-16k",
-    "maxToken": 16000,
-    "price": 0,
-    "prompt": ""
-  },
-  "CQModel": {
-    "model": "gpt-3.5-turbo-16k",
-    "functionCall": true,
-    "name": "GPT35-16k",
-    "maxToken": 16000,
-    "price": 0,
-    "prompt": ""
-  },
-  "QGModel": {
-    "model": "gpt-3.5-turbo",
-    "name": "GPT35-4k",
-    "maxToken": 4000,
-    "price": 0,
-    "prompt": "",
-    "functionCall": false
-  }
+  ]
 }
--- a/projects/app/package.json
+++ b/projects/app/package.json
@@ -1,6 +1,6 @@
 {
  "name": "app",
-  "version": "4.4.7",
+  "version": "4.5.0",
  "private": false,
  "scripts": {
    "dev": "next dev",
@@ -31,6 +31,7 @@
    "formidable": "^2.1.1",
    "framer-motion": "^9.0.6",
    "hyperdown": "^2.4.29",
+    "i18next": "^23.2.11",
    "immer": "^9.0.19",
    "js-cookie": "^3.0.5",
    "js-tiktoken": "^1.0.7",
@@ -43,7 +44,7 @@
    "multer": "1.4.5-lts.1",
    "nanoid": "^4.0.1",
    "next": "13.5.2",
-    "nextjs-cors": "^2.1.2",
+    "next-i18next": "^14.0.0",
    "nprogress": "^0.2.0",
    "papaparse": "^5.4.1",
    "pg": "^8.10.0",
@@ -52,6 +53,7 @@
    "react-day-picker": "^8.7.1",
    "react-dom": "18.2.0",
    "react-hook-form": "^7.43.1",
+    "react-i18next": "^13.0.2",
    "react-markdown": "^8.0.7",
    "react-syntax-highlighter": "^15.5.0",
    "reactflow": "^11.7.4",
@@ -62,10 +64,7 @@
    "request-ip": "^3.3.0",
    "sass": "^1.58.3",
    "timezones-list": "^3.0.2",
-    "zustand": "^4.3.5",
-    "i18next": "^23.2.11",
-    "react-i18next": "^13.0.2",
-    "next-i18next": "^14.0.0"
+    "zustand": "^4.3.5"
  },
  "devDependencies": {
    "@svgr/webpack": "^6.5.1",
@@ -76,6 +75,7 @@
    "@types/jsonwebtoken": "^9.0.3",
    "@types/lodash": "^4.14.191",
    "@types/multer": "^1.4.7",
+    "@types/node": "^20.8.5",
    "@types/papaparse": "^5.3.7",
    "@types/pg": "^8.6.6",
    "@types/react": "18.0.28",
--- a/projects/app/public/docs/versionIntro.md
+++ b/projects/app/public/docs/versionIntro.md
@@ -1,6 +1,10 @@
-### Fast GPT V4.4.7
+### Fast GPT V4.5.0

-1. 优化数据集管理，区分手动录入和标注，可追数据至某个文件，保留链接读取的原始链接。
-2. [使用文档](https://doc.fastgpt.run/docs/intro/)
-3. [点击查看高级编排介绍文档](https://doc.fastgpt.run/docs/workflow)
-4. [点击查看商业版](https://doc.fastgpt.run/docs/commercial/)
+1. 新增 - 升级 PgVector 插件，引入 HNSW 索引，极大加快的知识库搜索速度。
+2. 新增 - AI对话模块，增加【返回AI内容】选项，可控制 AI 的内容不直接返回浏览器。
+3. 优化 - TextSplitter，采用递归拆解法。
+4. 优化 - 高级编排 UX 性能
+5. 优化数据集管理，区分手动录入和标注，可追数据至某个文件，保留链接读取的原始链接。
+6. [使用文档](https://doc.fastgpt.run/docs/intro/)
+7. [点击查看高级编排介绍文档](https://doc.fastgpt.run/docs/workflow)
+8. [点击查看商业版](https://doc.fastgpt.run/docs/commercial/)
--- a/projects/app/public/locales/en/common.json
+++ b/projects/app/public/locales/en/common.json
@@ -39,7 +39,7 @@
    "My Apps": "My Apps",
    "Output Field Settings": "Output Field Settings",
    "Paste Config": "Paste Config",
-    "Quote Prompt Settings": "Quote Prompt Settings",
+    "AI Settings": "AI Settings",
    "Variable Key Repeat Tip": "Variable Key Repeat",
    "module": {
      "Custom Title Tip": "The title name is displayed during the conversation"
--- a/projects/app/public/locales/zh/common.json
+++ b/projects/app/public/locales/zh/common.json
@@ -39,7 +39,7 @@
    "My Apps": "我的应用",
    "Output Field Settings": "输出字段编辑",
    "Paste Config": "粘贴配置",
-    "Quote Prompt Settings": "引用提示词配置",
+    "AI Settings": "AI 高级配置",
    "Variable Key Repeat Tip": "变量 key 重复",
    "module": {
      "Custom Title Tip": "该标题名字会展示在对话过程中"
--- a/projects/app/src/components/ChatBox/utils.ts
+++ b/projects/app/src/components/ChatBox/utils.ts
@@ -1,6 +1,5 @@
 import { SystemInputEnum } from '@/constants/app';
 import { FlowModuleTypeEnum } from '@/constants/flow';
-import { getChatModel } from '@/service/utils/data';
 import { AppModuleItemType, VariableItemType } from '@/types/app';

 export const getGuideModule = (modules: AppModuleItemType[]) =>
@@ -23,11 +22,3 @@ export const splitGuideModule = (guideModules?: AppModuleItemType) => {
    questionGuide
  };
 };
-export const getChatModelNameList = (modules: AppModuleItemType[]): string[] => {
-  const chatModules = modules.filter((item) => item.flowType === FlowModuleTypeEnum.chatNode);
-  return chatModules
-    .map(
-      (item) => getChatModel(item.inputs.find((input) => input.key === 'model')?.value)?.name || ''
-    )
-    .filter((item) => item);
-};
--- a/projects/app/src/components/Markdown/index.tsx
+++ b/projects/app/src/components/Markdown/index.tsx
@@ -62,7 +62,9 @@ const Markdown = ({ source, isChatting = false }: { source: string; isChatting?:
    []
  );

-  const formatSource = source.replace(/\\n/g, '\n&nbsp;');
+  const formatSource = source
+    .replace(/\\n/g, '\n&nbsp;')
+    .replace(/(http[s]?:\/\/[^\s，。]+)([。，])/g, '$1 $2');

  return (
    <ReactMarkdown
--- a/projects/app/src/components/MyModal/index.tsx
+++ b/projects/app/src/components/MyModal/index.tsx
@@ -35,8 +35,6 @@ const MyModal = ({
    >
      <ModalOverlay />
      <ModalContent
-        display={'flex'}
-        flexDirection={'column'}
        w={w}
        minW={['90vw', '400px']}
        maxW={maxW}
@@ -46,7 +44,7 @@ const MyModal = ({
      >
        {!!title && <ModalHeader>{title}</ModalHeader>}
        {onClose && <ModalCloseButton />}
-        <Box overflow={'overlay'} h={'100%'}>
+        <Box overflow={'overlay'} h={'100%'} display={'flex'} flexDirection={'column'}>
          {children}
        </Box>
      </ModalContent>
--- a/projects/app/src/components/PromptTemplate/index.tsx
+++ b/projects/app/src/components/PromptTemplate/index.tsx
@@ -0,0 +1,64 @@
+import React, { useState } from 'react';
+import MyModal from '../MyModal';
+import { Box, Button, Grid, useTheme } from '@chakra-ui/react';
+import { PromptTemplateItem } from '@fastgpt/core/ai/type';
+import { ModalBody, ModalFooter } from '@chakra-ui/react';
+
+const PromptTemplate = ({
+  title,
+  templates,
+  onClose,
+  onSuccess
+}: {
+  title: string;
+  templates: PromptTemplateItem[];
+  onClose: () => void;
+  onSuccess: (e: string) => void;
+}) => {
+  const theme = useTheme();
+  const [selectTemplateTitle, setSelectTemplateTitle] = useState<PromptTemplateItem>();
+
+  return (
+    <MyModal isOpen title={title} onClose={onClose}>
+      <ModalBody w={'600px'}>
+        <Grid gridTemplateColumns={['1fr', '1fr 1fr']} gridGap={4}>
+          {templates.map((item) => (
+            <Box
+              key={item.title}
+              border={theme.borders.base}
+              py={2}
+              px={2}
+              borderRadius={'md'}
+              cursor={'pointer'}
+              {...(item.title === selectTemplateTitle?.title
+                ? {
+                    bg: 'myBlue.100'
+                  }
+                : {})}
+              onClick={() => setSelectTemplateTitle(item)}
+            >
+              <Box>{item.title}</Box>
+              <Box color={'myGray.600'} fontSize={'sm'} whiteSpace={'pre-wrap'}>
+                {item.value}
+              </Box>
+            </Box>
+          ))}
+        </Grid>
+      </ModalBody>
+      <ModalFooter>
+        <Button
+          disabled={!selectTemplateTitle}
+          onClick={() => {
+            if (!selectTemplateTitle) return;
+            onSuccess(selectTemplateTitle.value);
+            onClose();
+          }}
+        >
+          确认选择
+        </Button>
+      </ModalFooter>
+    </MyModal>
+  );
+};
+
+export default PromptTemplate;
--- a/projects/app/src/constants/app.ts
+++ b/projects/app/src/constants/app.ts
@@ -5,7 +5,8 @@ export enum SystemInputEnum {
  'switch' = 'switch', // a trigger switch
  'history' = 'history',
  'userChatInput' = 'userChatInput',
-  'questionGuide' = 'questionGuide'
+  'questionGuide' = 'questionGuide',
+  isResponseAnswerText = 'isResponseAnswerText'
 }
 export enum SystemOutputEnum {
  finish = 'finish'
--- a/projects/app/src/constants/flow/ModuleTemplate.ts
+++ b/projects/app/src/constants/flow/ModuleTemplate.ts
@@ -9,7 +9,7 @@ import {
 } from './index';
 import type { AppItemType } from '@/types/app';
 import type { FlowModuleTemplateType } from '@/types/core/app/flow';
-import { chatModelList } from '@/web/common/store/static';
+import { chatModelList, cqModelList } from '@/web/common/store/static';
 import {
  Input_Template_History,
  Input_Template_TFSwitch,
@@ -136,14 +136,14 @@ export const ChatModule: FlowModuleTemplateType = {
      key: 'model',
      type: FlowInputItemTypeEnum.selectChatModel,
      label: '对话模型',
-      value: chatModelList[0]?.model,
-      list: chatModelList.map((item) => ({ label: item.name, value: item.model })),
+      value: chatModelList?.[0]?.model,
+      customData: () => chatModelList,
      required: true,
      valueCheck: (val) => !!val
    },
    {
      key: 'temperature',
-      type: FlowInputItemTypeEnum.slider,
+      type: FlowInputItemTypeEnum.hidden,
      label: '温度',
      value: 0,
      min: 0,
@@ -156,20 +156,26 @@ export const ChatModule: FlowModuleTemplateType = {
    },
    {
      key: 'maxToken',
-      type: FlowInputItemTypeEnum.maxToken,
+      type: FlowInputItemTypeEnum.hidden,
      label: '回复上限',
-      value: chatModelList[0] ? chatModelList[0].contextMaxToken / 2 : 2000,
+      value: chatModelList?.[0] ? chatModelList[0].maxToken / 2 : 2000,
      min: 100,
-      max: chatModelList[0]?.contextMaxToken || 4000,
+      max: chatModelList?.[0]?.maxToken || 4000,
      step: 50,
      markList: [
        { label: '100', value: 100 },
        {
-          label: `${chatModelList[0]?.contextMaxToken || 4000}`,
-          value: chatModelList[0]?.contextMaxToken || 4000
+          label: `${chatModelList?.[0]?.maxToken || 4000}`,
+          value: chatModelList?.[0]?.maxToken || 4000
        }
      ]
    },
+    {
+      key: 'aiSettings',
+      type: FlowInputItemTypeEnum.aiSettings,
+      label: '',
+      connected: false
+    },
    {
      key: 'systemPrompt',
      type: FlowInputItemTypeEnum.textarea,
@@ -180,6 +186,13 @@ export const ChatModule: FlowModuleTemplateType = {
      placeholder: ChatModelSystemTip,
      value: ''
    },
+    {
+      key: SystemInputEnum.isResponseAnswerText,
+      type: FlowInputItemTypeEnum.hidden,
+      label: '返回AI内容',
+      valueType: FlowValueTypeEnum.boolean,
+      value: true
+    },
    {
      key: 'quoteTemplate',
      type: FlowInputItemTypeEnum.hidden,
@@ -196,7 +209,7 @@ export const ChatModule: FlowModuleTemplateType = {
    },
    {
      key: 'quoteQA',
-      type: FlowInputItemTypeEnum.quoteList,
+      type: FlowInputItemTypeEnum.target,
      label: '引用内容',
      description: "对象数组格式，结构：\n [{q:'问题',a:'回答'}]",
      valueType: FlowValueTypeEnum.kbQuote,
@@ -216,7 +229,7 @@ export const ChatModule: FlowModuleTemplateType = {
    },
    {
      key: TaskResponseKeyEnum.answerText,
-      label: '模型回复',
+      label: 'AI回复',
      description: '将在 stream 回复完毕后触发',
      valueType: FlowValueTypeEnum.string,
      type: FlowOutputItemTypeEnum.source,
@@ -330,12 +343,21 @@ export const ClassifyQuestionModule: FlowModuleTemplateType = {
  showStatus: true,
  inputs: [
    Input_Template_TFSwitch,
+    {
+      key: 'model',
+      type: FlowInputItemTypeEnum.selectChatModel,
+      label: '分类模型',
+      value: cqModelList?.[0]?.model,
+      customData: () => cqModelList,
+      required: true,
+      valueCheck: (val) => !!val
+    },
    {
      key: 'systemPrompt',
      type: FlowInputItemTypeEnum.textarea,
      valueType: FlowValueTypeEnum.string,
      value: '',
-      label: '系统提示词',
+      label: '背景知识',
      description:
        '你可以添加一些特定内容的介绍，从而更好的识别用户的问题类型。这个内容通常是给模型介绍一个它不知道的内容。',
      placeholder: '例如: \n1. Laf 是一个云函数开发平台……\n2. Sealos 是一个集群操作系统'
@@ -504,7 +526,7 @@ export const AppModule: FlowModuleTemplateType = {
    },
    {
      key: TaskResponseKeyEnum.answerText,
-      label: '模型回复',
+      label: 'AI回复',
      description: '将在应用完全结束后触发',
      valueType: FlowValueTypeEnum.string,
      type: FlowOutputItemTypeEnum.source,
@@ -757,7 +779,7 @@ export const appTemplates: (AppItemType & {
        outputs: [
          {
            key: 'answerText',
-            label: '模型回复',
+            label: 'AI回复',
            description: '直接响应，无需配置',
            type: 'hidden',
            targets: []
@@ -1094,7 +1116,7 @@ export const appTemplates: (AppItemType & {
        outputs: [
          {
            key: 'answerText',
-            label: '模型回复',
+            label: 'AI回复',
            description: '直接响应，无需配置',
            type: 'hidden',
            targets: []
@@ -1401,7 +1423,7 @@ export const appTemplates: (AppItemType & {
        outputs: [
          {
            key: 'answerText',
-            label: '模型回复',
+            label: 'AI回复',
            description: '将在 stream 回复完毕后触发',
            valueType: 'string',
            type: 'source',
@@ -1863,7 +1885,7 @@ export const appTemplates: (AppItemType & {
        outputs: [
          {
            key: 'answerText',
-            label: '模型回复',
+            label: 'AI回复',
            description: '将在 stream 回复完毕后触发',
            valueType: 'string',
            type: 'source',
--- a/projects/app/src/constants/flow/index.ts
+++ b/projects/app/src/constants/flow/index.ts
@@ -13,7 +13,7 @@ export enum FlowInputItemTypeEnum {
  chatInput = 'chatInput',
  selectApp = 'selectApp',
  // chat special input
-  quoteList = 'quoteList',
+  aiSettings = 'aiSettings',
  maxToken = 'maxToken',
  selectChatModel = 'selectChatModel',
  // dataset special input
--- a/projects/app/src/constants/model.ts
+++ b/projects/app/src/constants/model.ts
@@ -1,5 +1,98 @@
 import type { AppSchema } from '@/types/mongoSchema';
 import type { OutLinkEditType } from '@fastgpt/support/outLink/type.d';
+import type {
+  LLMModelItemType,
+  ChatModelItemType,
+  FunctionModelItemType,
+  VectorModelItemType
+} from '@/types/model';
+
+export const defaultChatModels: ChatModelItemType[] = [
+  {
+    model: 'gpt-3.5-turbo',
+    name: 'GPT35-4k',
+    price: 0,
+    maxToken: 4000,
+    quoteMaxToken: 2000,
+    maxTemperature: 1.2,
+    censor: false,
+    defaultSystemChatPrompt: ''
+  },
+  {
+    model: 'gpt-3.5-turbo-16k',
+    name: 'GPT35-16k',
+    maxToken: 16000,
+    price: 0,
+    quoteMaxToken: 8000,
+    maxTemperature: 1.2,
+    censor: false,
+    defaultSystemChatPrompt: ''
+  },
+  {
+    model: 'gpt-4',
+    name: 'GPT4-8k',
+    maxToken: 8000,
+    price: 0,
+    quoteMaxToken: 4000,
+    maxTemperature: 1.2,
+    censor: false,
+    defaultSystemChatPrompt: ''
+  }
+];
+export const defaultQAModels: LLMModelItemType[] = [
+  {
+    model: 'gpt-3.5-turbo-16k',
+    name: 'GPT35-16k',
+    maxToken: 16000,
+    price: 0
+  }
+];
+export const defaultCQModels: FunctionModelItemType[] = [
+  {
+    model: 'gpt-3.5-turbo-16k',
+    name: 'GPT35-16k',
+    maxToken: 16000,
+    price: 0,
+    functionCall: true,
+    functionPrompt: ''
+  },
+  {
+    model: 'gpt-4',
+    name: 'GPT4-8k',
+    maxToken: 8000,
+    price: 0,
+    functionCall: true,
+    functionPrompt: ''
+  }
+];
+export const defaultExtractModels: FunctionModelItemType[] = [
+  {
+    model: 'gpt-3.5-turbo-16k',
+    name: 'GPT35-16k',
+    maxToken: 16000,
+    price: 0,
+    functionCall: true,
+    functionPrompt: ''
+  }
+];
+export const defaultQGModels: LLMModelItemType[] = [
+  {
+    model: 'gpt-3.5-turbo',
+    name: 'GPT35-4K',
+    maxToken: 4000,
+    price: 0
+  }
+];
+
+export const defaultVectorModels: VectorModelItemType[] = [
+  {
+    model: 'text-embedding-ada-002',
+    name: 'Embedding-2',
+    price: 0,
+    defaultToken: 500,
+    maxToken: 3000
+  }
+];

 export const defaultApp: AppSchema = {
  _id: '',
--- a/projects/app/src/global/common/api/systemRes.d.ts
+++ b/projects/app/src/global/common/api/systemRes.d.ts
@@ -1,14 +1,17 @@
-import {
-  type QAModelItemType,
-  type ChatModelItemType,
-  type VectorModelItemType,
-  FunctionModelItemType
+import type {
+  ChatModelItemType,
+  FunctionModelItemType,
+  LLMModelItemType,
+  VectorModelItemType
 } from '@/types/model';
 import type { FeConfigsType } from '@fastgpt/common/type/index.d';

 export type InitDateResponse = {
  chatModels: ChatModelItemType[];
-  qaModel: QAModelItemType;
+  qaModels: LLMModelItemType[];
+  cqModels: FunctionModelItemType[];
+  extractModels: FunctionModelItemType[];
+  qgModels: LLMModelItemType[];
  vectorModels: VectorModelItemType[];
  feConfigs: FeConfigsType;
  priceMd: string;
--- a/projects/app/src/global/core/prompt/AIChat.ts
+++ b/projects/app/src/global/core/prompt/AIChat.ts
@@ -1,5 +1,23 @@
-export const defaultQuoteTemplate = `{instruction:"{{q}}",output:"{{a}}"}`;
-export const defaultQuotePrompt = `你的背景知识:
+import { PromptTemplateItem } from '@fastgpt/core/ai/type.d';
+
+export const Prompt_QuoteTemplateList: PromptTemplateItem[] = [
+  {
+    title: '标准模板',
+    desc: '包含 q 和 a 两个变量的标准模板',
+    value: `{instruction:"{{q}}",output:"{{a}}"}`
+  },
+  {
+    title: '全部变量',
+    desc: '包含 q 和 a 两个变量的标准模板',
+    value: `{instruction:"{{q}}",output:"{{a}}",source:"{{source}}",file_id:"{{file_id}}",index:"{{index}}"}`
+  }
+];
+
+export const Prompt_QuotePromptList: PromptTemplateItem[] = [
+  {
+    title: '标准模式',
+    desc: '',
+    value: `你的背景知识:
 """
 {{quote}}
 """
@@ -7,4 +25,19 @@ export const defaultQuotePrompt = `你的背景知识:
 1. 背景知识是最新的，其中 instruction 是相关介绍，output 是预期回答或补充。
 2. 使用背景知识回答问题。
 3. 背景知识无法满足问题时，你需严谨的回答问题。
-我的问题是:"{{question}}"`;
+我的问题是:"{{question}}"`
+  },
+  {
+    title: '严格模式',
+    desc: '',
+    value: `你的背景知识:
+"""
+{{quote}}
+"""
+对话要求：
+1. 背景知识是最新的，其中 instruction 是相关介绍，output 是预期回答或补充。
+2. 使用背景知识回答问题。
+3. 背景知识无法满足问题时，你需要回答：我不清楚关于xxx的内容。
+我的问题是:"{{question}}"`
+  }
+];
--- a/projects/app/src/pages/_error.tsx
+++ b/projects/app/src/pages/_error.tsx
@@ -32,8 +32,6 @@ function Error() {
 }

 export async function getServerSideProps(context: any) {
-  console.log('[render error]: ', context);
-
  return {
    props: { ...(await serviceSideProps(context)) }
  };
--- a/projects/app/src/pages/api/chat/chatTest.ts
+++ b/projects/app/src/pages/api/chat/chatTest.ts
@@ -3,7 +3,7 @@ import { connectToDatabase } from '@/service/mongo';
 import { authUser } from '@fastgpt/support/user/auth';
 import { sseErrRes } from '@/service/response';
 import { sseResponseEventEnum } from '@/constants/chat';
-import { sseResponse } from '@/service/utils/tools';
+import { responseWrite } from '@fastgpt/common/tools/stream';
 import { AppModuleItemType } from '@/types/app';
 import { dispatchModules } from '@/pages/api/v1/chat/completions';
 import { pushChatBill } from '@/service/common/bill/push';
@@ -59,12 +59,12 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
      detail: true
    });

-    sseResponse({
+    responseWrite({
      res,
      event: sseResponseEventEnum.answer,
      data: '[DONE]'
    });
-    sseResponse({
+    responseWrite({
      res,
      event: sseResponseEventEnum.appStreamResponse,
      data: JSON.stringify(responseData)
--- a/projects/app/src/pages/api/chat/init.ts
+++ b/projects/app/src/pages/api/chat/init.ts
@@ -6,7 +6,8 @@ import { authUser } from '@fastgpt/support/user/auth';
 import { ChatItemType } from '@/types/chat';
 import { authApp } from '@/service/utils/auth';
 import type { ChatSchema } from '@/types/mongoSchema';
-import { getChatModelNameList, getGuideModule } from '@/components/ChatBox/utils';
+import { getGuideModule } from '@/components/ChatBox/utils';
+import { getChatModelNameListByModules } from '@/service/core/app/module';
 import { TaskResponseKeyEnum } from '@/constants/chat';

 /* 初始化我的聊天框，需要身份验证 */
@@ -83,7 +84,7 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
        appId,
        app: {
          userGuideModule: getGuideModule(app.modules),
-          chatModels: getChatModelNameList(app.modules),
+          chatModels: getChatModelNameListByModules(app.modules),
          name: app.name,
          avatar: app.avatar,
          intro: app.intro,
--- a/projects/app/src/pages/api/common/bill/createTrainingBill.ts
+++ b/projects/app/src/pages/api/common/bill/createTrainingBill.ts
@@ -12,6 +12,8 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)

    const { userId } = await authUser({ req, authToken: true, authApiKey: true });

+    const qaModel = global.qaModels[0];
+
    const { _id } = await Bill.create({
      userId,
      appName: name,
@@ -25,7 +27,7 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
        },
        {
          moduleName: 'QA 拆分',
-          model: global.qaModel.name,
+          model: qaModel?.name,
          amount: 0,
          tokenLen: 0
        }
--- a/projects/app/src/pages/api/core/ai/agent/createQuestionGuide.ts
+++ b/projects/app/src/pages/api/core/ai/agent/createQuestionGuide.ts
@@ -4,7 +4,6 @@ import { connectToDatabase } from '@/service/mongo';
 import { authUser } from '@fastgpt/support/user/auth';
 import type { CreateQuestionGuideParams } from '@/global/core/api/aiReq.d';
 import { pushQuestionGuideBill } from '@/service/common/bill/push';
-import { defaultQGModel } from '@/pages/api/system/getInitData';
 import { createQuestionGuide } from '@fastgpt/core/ai/functions/createQuestionGuide';

 export default async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
@@ -23,9 +22,11 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse<
      throw new Error('user not found');
    }

+    const qgModel = global.qgModels[0];
+
    const { result, tokens } = await createQuestionGuide({
      messages,
-      model: (global.qgModel || defaultQGModel).model
+      model: qgModel.model
    });

    jsonRes(res, {
--- a/projects/app/src/pages/api/core/dataset/allDataset.ts
+++ b/projects/app/src/pages/api/core/dataset/allDataset.ts
@@ -3,7 +3,7 @@ import { jsonRes } from '@/service/response';
 import { connectToDatabase } from '@/service/mongo';
 import { MongoDataset } from '@fastgpt/core/dataset/schema';
 import { authUser } from '@fastgpt/support/user/auth';
-import { getVectorModel } from '@/service/utils/data';
+import { getVectorModel } from '@/service/core/ai/model';
 import type { DatasetsItemType } from '@/types/core/dataset';

 export default async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
--- a/projects/app/src/pages/api/core/dataset/data/delDataById.ts
+++ b/projects/app/src/pages/api/core/dataset/data/delDataById.ts
@@ -2,7 +2,7 @@ import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { authUser } from '@fastgpt/support/user/auth';
 import { PgClient } from '@/service/pg';
-import { withNextCors } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import { PgDatasetTableName } from '@/constants/plugin';
 import { connectToDatabase } from '@/service/mongo';

--- a/projects/app/src/pages/api/core/dataset/data/exportAll.ts
+++ b/projects/app/src/pages/api/core/dataset/data/exportAll.ts
@@ -8,7 +8,7 @@ import { findAllChildrenIds } from '../delete';
 import QueryStream from 'pg-query-stream';
 import { PgClient } from '@/service/pg';
 import { addLog } from '@/service/utils/tools';
-import { responseWriteController } from '@/service/common/stream';
+import { responseWriteController } from '@fastgpt/common/tools/stream';

 export default async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
  try {
--- a/projects/app/src/pages/api/core/dataset/data/insertData.ts
+++ b/projects/app/src/pages/api/core/dataset/data/insertData.ts
@@ -7,10 +7,10 @@ import { jsonRes } from '@/service/response';
 import { connectToDatabase } from '@/service/mongo';
 import { authDataset } from '@/service/utils/auth';
 import { authUser } from '@fastgpt/support/user/auth';
-import { withNextCors } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import { PgDatasetTableName } from '@/constants/plugin';
 import { insertData2Dataset, PgClient } from '@/service/pg';
-import { getVectorModel } from '@/service/utils/data';
+import { getVectorModel } from '@/service/core/ai/model';
 import { getVector } from '@/pages/api/openapi/plugin/vector';
 import { DatasetDataItemType } from '@/types/core/dataset/data';
 import { countPromptTokens } from '@/utils/common/tiktoken';
--- a/projects/app/src/pages/api/core/dataset/data/pushData.ts
+++ b/projects/app/src/pages/api/core/dataset/data/pushData.ts
@@ -5,15 +5,15 @@ import { connectToDatabase, TrainingData } from '@/service/mongo';
 import { MongoDataset } from '@fastgpt/core/dataset/schema';
 import { authUser } from '@fastgpt/support/user/auth';
 import { authDataset } from '@/service/utils/auth';
-import { withNextCors } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import { TrainingModeEnum } from '@/constants/plugin';
 import { startQueue } from '@/service/utils/tools';
-import { getVectorModel } from '@/service/utils/data';
 import { DatasetDataItemType } from '@/types/core/dataset/data';
 import { countPromptTokens } from '@/utils/common/tiktoken';
 import type { PushDataResponse } from '@/global/core/api/datasetRes.d';
 import type { PushDataProps } from '@/global/core/api/datasetReq.d';
 import { authFileIdValid } from '@/service/dataset/auth';
+import { getVectorModel } from '@/service/core/ai/model';

 const modeMap = {
  [TrainingModeEnum.index]: true,
@@ -71,7 +71,7 @@ export async function pushDataToKb({
      if (mode === TrainingModeEnum.index) {
        const vectorModel = (await MongoDataset.findById(kbId, 'vectorModel'))?.vectorModel;

-        return getVectorModel(vectorModel || global.vectorModels[0].model);
+        return getVectorModel(vectorModel);
      }
      return global.vectorModels[0];
    })()
@@ -79,7 +79,7 @@ export async function pushDataToKb({

  const modeMaxToken = {
    [TrainingModeEnum.index]: vectorModel.maxToken * 1.5,
-    [TrainingModeEnum.qa]: global.qaModel.maxToken * 0.8
+    [TrainingModeEnum.qa]: global.qaModels[0].maxToken * 0.8
  };

  // filter repeat or equal content
--- a/projects/app/src/pages/api/core/dataset/data/updateData.ts
+++ b/projects/app/src/pages/api/core/dataset/data/updateData.ts
@@ -2,7 +2,7 @@ import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { authUser } from '@fastgpt/support/user/auth';
 import { PgClient } from '@/service/pg';
-import { withNextCors } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import { connectToDatabase } from '@/service/mongo';
 import { MongoDataset } from '@fastgpt/core/dataset/schema';
 import { getVector } from '@/pages/api/openapi/plugin/vector';
--- a/projects/app/src/pages/api/core/dataset/detail.ts
+++ b/projects/app/src/pages/api/core/dataset/detail.ts
@@ -2,7 +2,7 @@ import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { connectToDatabase } from '@/service/mongo';
 import { authUser } from '@fastgpt/support/user/auth';
-import { getVectorModel } from '@/service/utils/data';
+import { getVectorModel } from '@/service/core/ai/model';
 import { MongoDataset } from '@fastgpt/core/dataset/schema';

 export default async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
--- a/projects/app/src/pages/api/core/dataset/list.ts
+++ b/projects/app/src/pages/api/core/dataset/list.ts
@@ -2,7 +2,7 @@ import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { connectToDatabase } from '@/service/mongo';
 import { authUser } from '@fastgpt/support/user/auth';
-import { getVectorModel } from '@/service/utils/data';
+import { getVectorModel } from '@/service/core/ai/model';
 import type { DatasetsItemType } from '@/types/core/dataset';
 import { DatasetTypeEnum } from '@fastgpt/core/dataset/constant';
 import { MongoDataset } from '@fastgpt/core/dataset/schema';
--- a/projects/app/src/pages/api/core/dataset/searchTest.ts
+++ b/projects/app/src/pages/api/core/dataset/searchTest.ts
@@ -2,7 +2,7 @@ import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { authUser } from '@fastgpt/support/user/auth';
 import { PgClient } from '@/service/pg';
-import { withNextCors } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import { getVector } from '../../openapi/plugin/vector';
 import { PgDatasetTableName } from '@/constants/plugin';
 import { MongoDataset } from '@fastgpt/core/dataset/schema';
--- a/projects/app/src/pages/api/openapi/plugin/vector.ts
+++ b/projects/app/src/pages/api/openapi/plugin/vector.ts
@@ -1,7 +1,7 @@
 import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { authBalanceByUid, authUser } from '@fastgpt/support/user/auth';
-import { withNextCors } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import { getAIApi } from '@fastgpt/core/ai/config';
 import { pushGenerateVectorBill } from '@/service/common/bill/push';
 import { connectToDatabase } from '@/service/mongo';
--- a/projects/app/src/pages/api/openapi/v1/chat/completions.ts
+++ b/projects/app/src/pages/api/openapi/v1/chat/completions.ts
@@ -1,5 +1,5 @@
 import type { NextApiRequest, NextApiResponse } from 'next';
-import { withNextCors } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import ChatCompletion from '@/pages/api/v1/chat/completions';

 export default withNextCors(async function handler(req: NextApiRequest, res: NextApiResponse) {
--- a/projects/app/src/pages/api/support/outLink/init.ts
+++ b/projects/app/src/pages/api/support/outLink/init.ts
@@ -6,8 +6,9 @@ import { MongoUser } from '@fastgpt/support/user/schema';
 import type { InitShareChatResponse } from '@/global/support/api/outLinkRes.d';
 import { authApp } from '@/service/utils/auth';
 import { HUMAN_ICON } from '@/constants/chat';
-import { getChatModelNameList, getGuideModule } from '@/components/ChatBox/utils';
+import { getGuideModule } from '@/components/ChatBox/utils';
 import { authShareChatInit } from '@fastgpt/support/outLink/auth';
+import { getChatModelNameListByModules } from '@/service/core/app/module';

 /* init share chat window */
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
@@ -51,7 +52,7 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
        userAvatar: user?.avatar || HUMAN_ICON,
        app: {
          userGuideModule: getGuideModule(app.modules),
-          chatModels: getChatModelNameList(app.modules),
+          chatModels: getChatModelNameListByModules(app.modules),
          name: app.name,
          avatar: app.avatar,
          intro: app.intro
--- a/projects/app/src/pages/api/system/getInitData.ts
+++ b/projects/app/src/pages/api/system/getInitData.ts
@@ -4,10 +4,23 @@ import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { readFileSync } from 'fs';
 import type { InitDateResponse } from '@/global/common/api/systemRes';
-import type { VectorModelItemType, FunctionModelItemType } from '@/types/model';
 import { formatPrice } from '@fastgpt/common/bill';
 import { getTikTokenEnc } from '@/utils/common/tiktoken';
 import { initHttpAgent } from '@fastgpt/core/init';
+import {
+  defaultChatModels,
+  defaultQAModels,
+  defaultCQModels,
+  defaultExtractModels,
+  defaultQGModels,
+  defaultVectorModels
+} from '@/constants/model';
+import {
+  ChatModelItemType,
+  FunctionModelItemType,
+  LLMModelItemType,
+  VectorModelItemType
+} from '@/types/model';

 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
  getInitConfig();
@@ -17,7 +30,10 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
    data: {
      feConfigs: global.feConfigs,
      chatModels: global.chatModels,
-      qaModel: global.qaModel,
+      qaModels: global.qaModels,
+      cqModels: global.cqModels,
+      extractModels: global.extractModels,
+      qgModels: global.qgModels,
      vectorModels: global.vectorModels,
      priceMd: global.priceMd,
      systemVersion: global.systemVersion || '0.0.0'
@@ -42,72 +58,6 @@ const defaultFeConfigs: FeConfigsType = {
  },
  scripts: []
 };
-const defaultChatModels = [
-  {
-    model: 'gpt-3.5-turbo',
-    name: 'GPT35-4k',
-    contextMaxToken: 4000,
-    quoteMaxToken: 2400,
-    maxTemperature: 1.2,
-    price: 0
-  },
-  {
-    model: 'gpt-3.5-turbo-16k',
-    name: 'GPT35-16k',
-    contextMaxToken: 16000,
-    quoteMaxToken: 8000,
-    maxTemperature: 1.2,
-    price: 0
-  },
-  {
-    model: 'gpt-4',
-    name: 'GPT4-8k',
-    contextMaxToken: 8000,
-    quoteMaxToken: 4000,
-    maxTemperature: 1.2,
-    price: 0
-  }
-];
-const defaultQAModel = {
-  model: 'gpt-3.5-turbo-16k',
-  name: 'GPT35-16k',
-  maxToken: 16000,
-  price: 0
-};
-export const defaultExtractModel: FunctionModelItemType = {
-  model: 'gpt-3.5-turbo-16k',
-  name: 'GPT35-16k',
-  maxToken: 16000,
-  price: 0,
-  prompt: '',
-  functionCall: true
-};
-export const defaultCQModel: FunctionModelItemType = {
-  model: 'gpt-3.5-turbo-16k',
-  name: 'GPT35-16k',
-  maxToken: 16000,
-  price: 0,
-  prompt: '',
-  functionCall: true
-};
-export const defaultQGModel: FunctionModelItemType = {
-  model: 'gpt-3.5-turbo',
-  name: 'FastAI-4k',
-  maxToken: 4000,
-  price: 1.5,
-  prompt: '',
-  functionCall: false
-};
-
-const defaultVectorModels: VectorModelItemType[] = [
-  {
-    model: 'text-embedding-ada-002',
-    name: 'Embedding-2',
-    price: 0,
-    defaultToken: 500,
-    maxToken: 3000
-  }
-];

 export function initGlobal() {
  // init tikToken
@@ -127,7 +77,16 @@ export function getInitConfig() {

    const filename =
      process.env.NODE_ENV === 'development' ? 'data/config.local.json' : '/app/data/config.json';
-    const res = JSON.parse(readFileSync(filename, 'utf-8'));
+    const res = JSON.parse(readFileSync(filename, 'utf-8')) as {
+      FeConfig: FeConfigsType;
+      SystemParams: SystemEnvType;
+      ChatModels: ChatModelItemType[];
+      QAModels: LLMModelItemType[];
+      CQModels: FunctionModelItemType[];
+      ExtractModels: FunctionModelItemType[];
+      QGModels: LLMModelItemType[];
+      VectorModels: VectorModelItemType[];
+    };

    console.log(`System Version: ${global.systemVersion}`);

@@ -137,11 +96,13 @@ export function getInitConfig() {
      ? { ...defaultSystemEnv, ...res.SystemParams }
      : defaultSystemEnv;
    global.feConfigs = res.FeConfig ? { ...defaultFeConfigs, ...res.FeConfig } : defaultFeConfigs;
+
    global.chatModels = res.ChatModels || defaultChatModels;
-    global.qaModel = res.QAModel || defaultQAModel;
-    global.extractModel = res.ExtractModel || defaultExtractModel;
-    global.cqModel = res.CQModel || defaultCQModel;
-    global.qgModel = res.QGModel || defaultQGModel;
+    global.qaModels = res.QAModels || defaultQAModels;
+    global.cqModels = res.CQModels || defaultCQModels;
+    global.extractModels = res.ExtractModels || defaultExtractModels;
+    global.qgModels = res.QGModels || defaultQGModels;
+
    global.vectorModels = res.VectorModels || defaultVectorModels;
  } catch (error) {
    setDefaultData();
@@ -152,13 +113,27 @@ export function getInitConfig() {
 export function setDefaultData() {
  global.systemEnv = defaultSystemEnv;
  global.feConfigs = defaultFeConfigs;
+
  global.chatModels = defaultChatModels;
-  global.qaModel = defaultQAModel;
+  global.qaModels = defaultQAModels;
+  global.cqModels = defaultCQModels;
+  global.extractModels = defaultExtractModels;
+  global.qgModels = defaultQGModels;
+
  global.vectorModels = defaultVectorModels;
-  global.extractModel = defaultExtractModel;
-  global.cqModel = defaultCQModel;
-  global.qgModel = defaultQGModel;
  global.priceMd = '';
+
+  console.log('use default config');
+  console.log({
+    feConfigs: defaultFeConfigs,
+    systemEnv: defaultSystemEnv,
+    chatModels: defaultChatModels,
+    qaModels: defaultQAModels,
+    cqModels: defaultCQModels,
+    extractModels: defaultExtractModels,
+    qgModels: defaultQGModels,
+    vectorModels: defaultVectorModels
+  });
 }

 export function getSystemVersion() {
@@ -187,10 +162,18 @@ ${global.vectorModels
 ${global.chatModels
  ?.map((item) => `| 对话-${item.name} | ${formatPrice(item.price, 1000)} |`)
  .join('\n')}
-| 文件QA拆分 | ${formatPrice(global.qaModel?.price, 1000)} |
-| 高级编排 - 问题分类 | ${formatPrice(global.cqModel?.price, 1000)} |
-| 高级编排 - 内容提取 | ${formatPrice(global.extractModel?.price, 1000)} |
-| 下一步指引 | ${formatPrice(global.qgModel?.price, 1000)} |
+${global.qaModels
+  ?.map((item) => `| 文件QA拆分-${item.name} | ${formatPrice(item.price, 1000)} |`)
+  .join('\n')}
+${global.cqModels
+  ?.map((item) => `| 问题分类-${item.name} | ${formatPrice(item.price, 1000)} |`)
+  .join('\n')}
+${global.extractModels
+  ?.map((item) => `| 内容提取-${item.name} | ${formatPrice(item.price, 1000)} |`)
+  .join('\n')}
+${global.qgModels
+  ?.map((item) => `| 下一步指引-${item.name} | ${formatPrice(item.price, 1000)} |`)
+  .join('\n')}
 `;
  console.log(global.priceMd);
 }
--- a/projects/app/src/pages/api/user/account/loginByPassword.ts
+++ b/projects/app/src/pages/api/user/account/loginByPassword.ts
@@ -2,8 +2,8 @@
 import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { MongoUser } from '@fastgpt/support/user/schema';
-import { setCookie } from '@/service/utils/tools';
-import { generateToken } from '@fastgpt/support/user/tools';
+import { setCookie } from '@fastgpt/support/user/auth';
+import { generateToken } from '@fastgpt/support/user/auth';
 import { connectToDatabase } from '@/service/mongo';

 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
--- a/projects/app/src/pages/api/user/account/loginout.ts
+++ b/projects/app/src/pages/api/user/account/loginout.ts
@@ -1,7 +1,7 @@
 // Next.js API route support: https://nextjs.org/docs/api-routes/introduction
 import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
-import { clearCookie } from '@/service/utils/tools';
+import { clearCookie } from '@fastgpt/support/user/auth';

 export default async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
  try {
--- a/projects/app/src/pages/api/v1/chat/completions.ts
+++ b/projects/app/src/pages/api/v1/chat/completions.ts
@@ -3,7 +3,8 @@ import { authApp } from '@/service/utils/auth';
 import { authUser } from '@fastgpt/support/user/auth';
 import { AuthUserTypeEnum } from '@fastgpt/support/user/auth';
 import { sseErrRes, jsonRes } from '@/service/response';
-import { addLog, withNextCors } from '@/service/utils/tools';
+import { addLog } from '@/service/utils/tools';
+import { withNextCors } from '@fastgpt/common/tools/nextjs';
 import { ChatRoleEnum, ChatSourceEnum, sseResponseEventEnum } from '@/constants/chat';
 import {
  dispatchHistory,
@@ -21,7 +22,7 @@ import type { MessageItemType } from '@/types/core/chat/type';
 import { gptMessage2ChatType, textAdaptGptResponse } from '@/utils/adapt';
 import { getChatHistory } from './getHistory';
 import { saveChat } from '@/service/utils/chat/saveChat';
-import { sseResponse } from '@/service/utils/tools';
+import { responseWrite } from '@fastgpt/common/tools/stream';
 import { TaskResponseKeyEnum } from '@/constants/chat';
 import { FlowModuleTypeEnum, initModuleType } from '@/constants/flow';
 import { AppModuleItemType, RunningModuleItemType } from '@/types/app';
@@ -217,7 +218,7 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
    const feResponseData = isOwner ? responseData : selectShareResponse({ responseData });

    if (stream) {
-      sseResponse({
+      responseWrite({
        res,
        event: detail ? sseResponseEventEnum.answer : undefined,
        data: textAdaptGptResponse({
@@ -225,14 +226,14 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
          finish_reason: 'stop'
        })
      });
-      sseResponse({
+      responseWrite({
        res,
        event: detail ? sseResponseEventEnum.answer : undefined,
        data: '[DONE]'
      });

      if (responseDetail && detail) {
-        sseResponse({
+        responseWrite({
          res,
          event: sseResponseEventEnum.appStreamResponse,
          data: JSON.stringify(feResponseData)
@@ -323,13 +324,16 @@ export async function dispatchModules({
  let chatAnswerText = ''; // AI answer
  let runningTime = Date.now();

-  function pushStore({
-    answerText = '',
-    responseData
-  }: {
-    answerText?: string;
-    responseData?: ChatHistoryItemResType | ChatHistoryItemResType[];
-  }) {
+  function pushStore(
+    { inputs = [] }: RunningModuleItemType,
+    {
+      answerText = '',
+      responseData
+    }: {
+      answerText?: string;
+      responseData?: ChatHistoryItemResType | ChatHistoryItemResType[];
+    }
+  ) {
    const time = Date.now();
    if (responseData) {
      if (Array.isArray(responseData)) {
@@ -342,7 +346,12 @@ export async function dispatchModules({
      }
    }
    runningTime = time;
-    chatAnswerText += answerText;
+
+    const isResponseAnswerText =
+      inputs.find((item) => item.key === SystemInputEnum.isResponseAnswerText)?.value ?? true;
+    if (isResponseAnswerText) {
+      chatAnswerText += answerText;
+    }
  }
  function moduleInput(
    module: RunningModuleItemType,
@@ -376,7 +385,7 @@ export async function dispatchModules({
    module: RunningModuleItemType,
    result: Record<string, any> = {}
  ): Promise<any> {
-    pushStore(result);
+    pushStore(module, result);
    return Promise.all(
      module.outputs.map((outputItem) => {
        if (result[outputItem.key] === undefined) return;
@@ -505,7 +514,7 @@ export function responseStatus({
  name?: string;
 }) {
  if (!name) return;
-  sseResponse({
+  responseWrite({
    res,
    event: sseResponseEventEnum.moduleStatus,
    data: JSON.stringify({
--- a/projects/app/src/pages/app/detail/components/AIChatSettingsModal.tsx
+++ b/projects/app/src/pages/app/detail/components/AIChatSettingsModal.tsx
@@ -1,4 +1,4 @@
-import React from 'react';
+import React, { useMemo, useState } from 'react';
 import MyModal from '@/components/MyModal';
 import { useTranslation } from 'react-i18next';
 import { EditFormType } from '@/utils/app';
@@ -11,43 +11,65 @@ import {
  Link,
  ModalBody,
  ModalFooter,
+  Switch,
  Textarea
 } from '@chakra-ui/react';
 import MyTooltip from '@/components/MyTooltip';
 import { QuestionOutlineIcon } from '@chakra-ui/icons';
-import { defaultQuotePrompt, defaultQuoteTemplate } from '@/global/core/prompt/AIChat';
-import { feConfigs } from '@/web/common/store/static';
+import { Prompt_QuotePromptList, Prompt_QuoteTemplateList } from '@/global/core/prompt/AIChat';
+import { chatModelList, feConfigs } from '@/web/common/store/static';
+import MySlider from '@/components/Slider';
+import { SystemInputEnum } from '@/constants/app';
+import dynamic from 'next/dynamic';
+import { PromptTemplateItem } from '@fastgpt/core/ai/type';
+
+const PromptTemplate = dynamic(() => import('@/components/PromptTemplate'));

 const AIChatSettingsModal = ({
+  isAdEdit,
  onClose,
  onSuccess,
  defaultData
 }: {
+  isAdEdit?: boolean;
  onClose: () => void;
  onSuccess: (e: EditFormType['chatModel']) => void;
  defaultData: EditFormType['chatModel'];
 }) => {
  const { t } = useTranslation();
+  const [refresh, setRefresh] = useState(false);

-  const { register, handleSubmit } = useForm({
+  const { register, handleSubmit, getValues, setValue } = useForm({
    defaultValues: defaultData
  });

+  const [selectTemplateData, setSelectTemplateData] = useState<{
+    title: string;
+    key: 'quoteTemplate' | 'quotePrompt';
+    templates: PromptTemplateItem[];
+  }>();
+
+  const tokenLimit = useMemo(() => {
+    return chatModelList.find((item) => item.model === getValues('model'))?.maxToken || 4000;
+  }, [getValues, refresh]);
+
  const LabelStyles: BoxProps = {
-    fontWeight: 'bold',
-    mb: 1,
    fontSize: ['sm', 'md']
  };
+  const selectTemplateBtn: BoxProps = {
+    color: 'myBlue.600',
+    cursor: 'pointer'
+  };

  return (
    <MyModal
      isOpen
      title={
        <Flex alignItems={'flex-end'}>
-          {t('app.Quote Prompt Settings')}
+          {t('app.AI Settings')}
          {feConfigs?.show_doc && (
            <Link
-              href={'https://doc.fastgpt.run/docs/use-cases/prompt/'}
+              href={'https://doc.fastgpt.run/docs/use-cases/ai_settings/'}
              target={'_blank'}
              ml={1}
              textDecoration={'underline'}
@@ -59,39 +81,134 @@ const AIChatSettingsModal = ({
          )}
        </Flex>
      }
+      isCentered
      w={'700px'}
+      h={['90vh', 'auto']}
    >
-      <ModalBody>
+      <ModalBody flex={['1 0 0', 'auto']} overflowY={'auto'}>
+        {isAdEdit && (
+          <Flex alignItems={'center'}>
+            <Box {...LabelStyles} w={'80px'}>
+              返回AI内容
+            </Box>
+            <Box flex={1} ml={'10px'}>
+              <Switch
+                isChecked={getValues(SystemInputEnum.isResponseAnswerText)}
+                size={'lg'}
+                onChange={(e) => {
+                  const value = e.target.checked;
+                  setValue(SystemInputEnum.isResponseAnswerText, value);
+                  setRefresh((state) => !state);
+                }}
+              />
+            </Box>
+          </Flex>
+        )}
+        <Flex alignItems={'center'} mb={10} mt={isAdEdit ? 8 : 5}>
+          <Box {...LabelStyles} mr={2} w={'80px'}>
+            温度
+          </Box>
+          <Box flex={1} ml={'10px'}>
+            <MySlider
+              markList={[
+                { label: '严谨', value: 0 },
+                { label: '发散', value: 10 }
+              ]}
+              width={'95%'}
+              min={0}
+              max={10}
+              value={getValues('temperature')}
+              onChange={(e) => {
+                setValue('temperature', e);
+                setRefresh(!refresh);
+              }}
+            />
+          </Box>
+        </Flex>
+        <Flex alignItems={'center'} mt={12} mb={10}>
+          <Box {...LabelStyles} mr={2} w={'80px'}>
+            回复上限
+          </Box>
+          <Box flex={1} ml={'10px'}>
+            <MySlider
+              markList={[
+                { label: '100', value: 100 },
+                { label: `${tokenLimit}`, value: tokenLimit }
+              ]}
+              width={'95%'}
+              min={100}
+              max={tokenLimit}
+              step={50}
+              value={getValues('maxToken')}
+              onChange={(val) => {
+                setValue('maxToken', val);
+                setRefresh(!refresh);
+              }}
+            />
+          </Box>
+        </Flex>
        <Box>
-          <Box {...LabelStyles}>
+          <Flex {...LabelStyles} mb={1}>
            引用内容模板
            <MyTooltip
-              label={t('template.Quote Content Tip', { default: defaultQuoteTemplate })}
+              label={t('template.Quote Content Tip', {
+                default: Prompt_QuoteTemplateList[0].value
+              })}
              forceShow
            >
              <QuestionOutlineIcon display={['none', 'inline']} ml={1} />
            </MyTooltip>
-          </Box>
+            <Box flex={1} />
+            <Box
+              {...selectTemplateBtn}
+              onClick={() =>
+                setSelectTemplateData({
+                  title: '选择引用内容模板',
+                  key: 'quoteTemplate',
+                  templates: Prompt_QuoteTemplateList
+                })
+              }
+            >
+              选择模板
+            </Box>
+          </Flex>
          <Textarea
-            rows={4}
-            placeholder={t('template.Quote Content Tip', { default: defaultQuoteTemplate }) || ''}
+            rows={6}
+            placeholder={
+              t('template.Quote Content Tip', { default: Prompt_QuoteTemplateList[0].value }) || ''
+            }
            borderColor={'myGray.100'}
            {...register('quoteTemplate')}
          />
        </Box>
        <Box mt={4}>
-          <Box {...LabelStyles}>
+          <Flex {...LabelStyles} mb={1}>
            引用内容提示词
            <MyTooltip
-              label={t('template.Quote Prompt Tip', { default: defaultQuotePrompt })}
+              label={t('template.Quote Prompt Tip', { default: Prompt_QuotePromptList[0].value })}
              forceShow
            >
              <QuestionOutlineIcon display={['none', 'inline']} ml={1} />
            </MyTooltip>
-          </Box>
+            <Box flex={1} />
+            <Box
+              {...selectTemplateBtn}
+              onClick={() =>
+                setSelectTemplateData({
+                  title: '选择引用提示词模板',
+                  key: 'quotePrompt',
+                  templates: Prompt_QuotePromptList
+                })
+              }
+            >
+              选择模板
+            </Box>
+          </Flex>
          <Textarea
-            rows={6}
-            placeholder={t('template.Quote Prompt Tip', { default: defaultQuotePrompt }) || ''}
+            rows={11}
+            placeholder={
+              t('template.Quote Prompt Tip', { default: Prompt_QuotePromptList[0].value }) || ''
+            }
            borderColor={'myGray.100'}
            {...register('quotePrompt')}
          />
@@ -105,6 +222,14 @@ const AIChatSettingsModal = ({
          {t('Confirm')}
        </Button>
      </ModalFooter>
+      {!!selectTemplateData && (
+        <PromptTemplate
+          title={selectTemplateData.title}
+          templates={selectTemplateData.templates}
+          onClose={() => setSelectTemplateData(undefined)}
+          onSuccess={(e) => setValue(selectTemplateData.key, e)}
+        />
+      )}
    </MyModal>
  );
 };
--- a/projects/app/src/pages/app/detail/components/AdEdit/components/Header.tsx
+++ b/projects/app/src/pages/app/detail/components/AdEdit/components/Header.tsx
@@ -0,0 +1,229 @@
+import React, { useCallback, useRef, useState } from 'react';
+import { Box, Flex, IconButton, useTheme, useDisclosure } from '@chakra-ui/react';
+import { SmallCloseIcon } from '@chakra-ui/icons';
+import { FlowInputItemTypeEnum } from '@/constants/flow';
+import { FlowOutputTargetItemType } from '@/types/core/app/flow';
+import { AppModuleItemType } from '@/types/app';
+import { useRequest } from '@/web/common/hooks/useRequest';
+import type { AppSchema } from '@/types/mongoSchema';
+import { useUserStore } from '@/web/support/store/user';
+import { useTranslation } from 'next-i18next';
+import { useCopyData } from '@/web/common/hooks/useCopyData';
+import { AppTypeEnum } from '@/constants/app';
+import dynamic from 'next/dynamic';
+
+import MyIcon from '@/components/Icon';
+import MyTooltip from '@/components/MyTooltip';
+import ChatTest, { type ChatTestComponentRef } from './ChatTest';
+import { useFlowStore } from './Provider';
+
+const ImportSettings = dynamic(() => import('./ImportSettings'));
+
+type Props = { app: AppSchema; onCloseSettings: () => void };
+
+const RenderHeaderContainer = React.memo(function RenderHeaderContainer({
+  app,
+  ChatTestRef,
+  testModules,
+  setTestModules,
+  onCloseSettings
+}: Props & {
+  ChatTestRef: React.RefObject<ChatTestComponentRef>;
+  testModules?: AppModuleItemType[];
+  setTestModules: React.Dispatch<AppModuleItemType[] | undefined>;
+}) {
+  const theme = useTheme();
+  const { t } = useTranslation();
+  const { copyData } = useCopyData();
+  const { isOpen: isOpenImport, onOpen: onOpenImport, onClose: onCloseImport } = useDisclosure();
+  const { updateAppDetail } = useUserStore();
+
+  const { nodes, edges, onFixView } = useFlowStore();
+
+  const flow2AppModules = useCallback(() => {
+    const modules: AppModuleItemType[] = nodes.map((item) => ({
+      moduleId: item.data.moduleId,
+      name: item.data.name,
+      flowType: item.data.flowType,
+      showStatus: item.data.showStatus,
+      position: item.position,
+      inputs: item.data.inputs.map((item) => ({
+        ...item,
+        connected: item.connected ?? item.type !== FlowInputItemTypeEnum.target
+      })),
+      outputs: item.data.outputs.map((item) => ({
+        ...item,
+        targets: [] as FlowOutputTargetItemType[]
+      }))
+    }));
+
+    // update inputs and outputs
+    modules.forEach((module) => {
+      module.inputs.forEach((input) => {
+        input.connected =
+          input.connected ||
+          !!edges.find(
+            (edge) => edge.target === module.moduleId && edge.targetHandle === input.key
+          );
+      });
+      module.outputs.forEach((output) => {
+        output.targets = edges
+          .filter(
+            (edge) =>
+              edge.source === module.moduleId &&
+              edge.sourceHandle === output.key &&
+              edge.targetHandle
+          )
+          .map((edge) => ({
+            moduleId: edge.target,
+            key: edge.targetHandle || ''
+          }));
+      });
+    });
+    return modules;
+  }, [edges, nodes]);
+
+  const { mutate: onclickSave, isLoading } = useRequest({
+    mutationFn: () => {
+      const modules = flow2AppModules();
+      // check required connect
+      for (let i = 0; i < modules.length; i++) {
+        const item = modules[i];
+        if (item.inputs.find((input) => input.required && !input.connected)) {
+          return Promise.reject(`【${item.name}】存在未连接的必填输入`);
+        }
+        if (item.inputs.find((input) => input.valueCheck && !input.valueCheck(input.value))) {
+          return Promise.reject(`【${item.name}】存在为填写的必填项`);
+        }
+      }
+
+      return updateAppDetail(app._id, {
+        modules: modules,
+        type: AppTypeEnum.advanced
+      });
+    },
+    successToast: '保存配置成功',
+    errorToast: '保存配置异常',
+    onSuccess() {
+      ChatTestRef.current?.resetChatTest();
+    }
+  });
+
+  return (
+    <>
+      <Flex
+        py={3}
+        px={[2, 5, 8]}
+        borderBottom={theme.borders.base}
+        alignItems={'center'}
+        userSelect={'none'}
+      >
+        <MyTooltip label={'返回'} offset={[10, 10]}>
+          <IconButton
+            size={'sm'}
+            icon={<MyIcon name={'back'} w={'14px'} />}
+            borderRadius={'md'}
+            borderColor={'myGray.300'}
+            variant={'base'}
+            aria-label={''}
+            onClick={() => {
+              onCloseSettings();
+              onFixView();
+            }}
+          />
+        </MyTooltip>
+        <Box ml={[3, 6]} fontSize={['md', '2xl']} flex={1}>
+          {app.name}
+        </Box>
+
+        <MyTooltip label={t('app.Import Configs')}>
+          <IconButton
+            mr={[3, 6]}
+            icon={<MyIcon name={'importLight'} w={['14px', '16px']} />}
+            borderRadius={'lg'}
+            variant={'base'}
+            aria-label={'save'}
+            onClick={onOpenImport}
+          />
+        </MyTooltip>
+        <MyTooltip label={t('app.Export Configs')}>
+          <IconButton
+            mr={[3, 6]}
+            icon={<MyIcon name={'export'} w={['14px', '16px']} />}
+            borderRadius={'lg'}
+            variant={'base'}
+            aria-label={'save'}
+            onClick={() =>
+              copyData(
+                JSON.stringify(flow2AppModules(), null, 2),
+                t('app.Export Config Successful')
+              )
+            }
+          />
+        </MyTooltip>
+
+        {testModules ? (
+          <IconButton
+            mr={[3, 6]}
+            icon={<SmallCloseIcon fontSize={'25px'} />}
+            variant={'base'}
+            color={'myGray.600'}
+            borderRadius={'lg'}
+            aria-label={''}
+            onClick={() => setTestModules(undefined)}
+          />
+        ) : (
+          <MyTooltip label={'测试对话'}>
+            <IconButton
+              mr={[3, 6]}
+              icon={<MyIcon name={'chat'} w={['14px', '16px']} />}
+              borderRadius={'lg'}
+              aria-label={'save'}
+              variant={'base'}
+              onClick={() => {
+                setTestModules(flow2AppModules());
+              }}
+            />
+          </MyTooltip>
+        )}
+
+        <MyTooltip label={'保存配置'}>
+          <IconButton
+            icon={<MyIcon name={'save'} w={['14px', '16px']} />}
+            borderRadius={'lg'}
+            isLoading={isLoading}
+            aria-label={'save'}
+            onClick={onclickSave}
+          />
+        </MyTooltip>
+      </Flex>
+      {isOpenImport && <ImportSettings onClose={onCloseImport} />}
+    </>
+  );
+});
+
+const Header = (props: Props) => {
+  const { app } = props;
+  const ChatTestRef = useRef<ChatTestComponentRef>(null);
+
+  const [testModules, setTestModules] = useState<AppModuleItemType[]>();
+
+  return (
+    <>
+      <RenderHeaderContainer
+        {...props}
+        ChatTestRef={ChatTestRef}
+        testModules={testModules}
+        setTestModules={setTestModules}
+      />
+      <ChatTest
+        ref={ChatTestRef}
+        modules={testModules}
+        app={app}
+        onClose={() => setTestModules(undefined)}
+      />
+    </>
+  );
+};
+
+export default React.memo(Header);
--- a/projects/app/src/pages/app/detail/components/AdEdit/components/Nodes/NodeChat.tsx
+++ b/projects/app/src/pages/app/detail/components/AdEdit/components/Nodes/NodeChat.tsx
@@ -1,4 +1,4 @@
-import React, { useMemo } from 'react';
+import React from 'react';
 import { NodeProps } from 'reactflow';
 import NodeCard from '../modules/NodeCard';
 import { FlowModuleItemType } from '@/types/core/app/flow';
@@ -7,11 +7,8 @@ import Container from '../modules/Container';
 import RenderInput from '../render/RenderInput';
 import RenderOutput from '../render/RenderOutput';

-import { useFlowStore } from '../Provider';
-
 const NodeChat = ({ data }: NodeProps<FlowModuleItemType>) => {
  const { moduleId, inputs, outputs } = data;
-  const { onChangeNode } = useFlowStore();

  return (
    <NodeCard minW={'400px'} {...data}>
--- a/projects/app/src/pages/app/detail/components/AdEdit/components/Provider.tsx
+++ b/projects/app/src/pages/app/detail/components/AdEdit/components/Provider.tsx
@@ -5,14 +5,11 @@ import {
  type EdgeChange,
  useNodesState,
  useEdgesState,
-  XYPosition,
-  useViewport,
  Connection,
  addEdge
 } from 'reactflow';
 import type {
  FlowModuleItemType,
-  FlowModuleTemplateType,
  FlowOutputTargetItemType,
  FlowModuleItemChangeProps
 } from '@/types/core/app/flow';
@@ -44,7 +41,6 @@ export type useFlowStoreType = {
  setEdges: Dispatch<SetStateAction<Edge<any>[]>>;
  onEdgesChange: OnChange<EdgeChange>;
  onFixView: () => void;
-  onAddNode: (e: { template: FlowModuleTemplateType; position: XYPosition }) => void;
  onDelNode: (nodeId: string) => void;
  onChangeNode: (e: FlowModuleItemChangeProps) => void;
  onCopyNode: (nodeId: string) => void;
@@ -80,9 +76,7 @@ const StateContext = createContext<useFlowStoreType>({
  onFixView: function (): void {
    return;
  },
-  onAddNode: function (e: { template: FlowModuleTemplateType; position: XYPosition }): void {
-    return;
-  },
+
  onDelNode: function (nodeId: string): void {
    return;
  },
@@ -117,7 +111,6 @@ export const FlowProvider = ({ appId, children }: { appId: string; children: Rea
  const { toast } = useToast();
  const [nodes = [], setNodes, onNodesChange] = useNodesState<FlowModuleItemType>([]);
  const [edges, setEdges, onEdgesChange] = useEdgesState([]);
-  const { x, y, zoom } = useViewport();

  const onFixView = useCallback(() => {
    const btn = document.querySelector('.react-flow__controls-fitview') as HTMLButtonElement;
@@ -205,27 +198,6 @@ export const FlowProvider = ({ appId, children }: { appId: string; children: Rea
    [nodes, onDelConnect, setEdges, t, toast]
  );

-  const onAddNode = useCallback(
-    ({ template, position }: { template: FlowModuleTemplateType; position: XYPosition }) => {
-      if (!reactFlowWrapper.current) return;
-      const reactFlowBounds = reactFlowWrapper.current.getBoundingClientRect();
-      const mouseX = (position.x - reactFlowBounds.left - x) / zoom - 100;
-      const mouseY = (position.y - reactFlowBounds.top - y) / zoom;
-      setNodes((state) =>
-        state.concat(
-          appModule2FlowNode({
-            item: {
-              ...template,
-              moduleId: nanoid(),
-              position: { x: mouseX, y: mouseY }
-            }
-          })
-        )
-      );
-    },
-    [setNodes, x, y, zoom]
-  );
-
  const onDelNode = useCallback(
    (nodeId: string) => {
      setNodes((state) => state.filter((item) => item.id !== nodeId));
@@ -338,7 +310,6 @@ export const FlowProvider = ({ appId, children }: { appId: string; children: Rea
    setEdges,
    onEdgesChange,
    onFixView,
-    onAddNode,
    onDelNode,
    onChangeNode,
    onCopyNode,
--- a/projects/app/src/pages/app/detail/components/AdEdit/components/TemplateList.tsx
+++ b/projects/app/src/pages/app/detail/components/AdEdit/components/TemplateList.tsx
@@ -1,24 +1,20 @@
-import React, { useMemo } from 'react';
+import React, { useCallback, useMemo } from 'react';
 import { Box, Flex } from '@chakra-ui/react';
 import { ModuleTemplates } from '@/constants/flow/ModuleTemplate';
-import { FlowModuleItemType, FlowModuleTemplateType } from '@/types/core/app/flow';
-import type { Node } from 'reactflow';
+import { FlowModuleTemplateType } from '@/types/core/app/flow';
+import { useViewport, XYPosition } from 'reactflow';
 import { useGlobalStore } from '@/web/common/store/global';
 import Avatar from '@/components/Avatar';
 import { FlowModuleTypeEnum } from '@/constants/flow';
 import { useFlowStore } from './Provider';
+import { customAlphabet } from 'nanoid';
+import { appModule2FlowNode } from '@/utils/adapt';
+const nanoid = customAlphabet('abcdefghijklmnopqrstuvwxyz1234567890', 6);

-const ModuleTemplateList = ({
-  nodes,
-  isOpen,
-  onClose
-}: {
-  nodes?: Node<FlowModuleItemType>[];
-  isOpen: boolean;
-  onClose: () => void;
-}) => {
-  const { onAddNode } = useFlowStore();
+const ModuleTemplateList = ({ isOpen, onClose }: { isOpen: boolean; onClose: () => void }) => {
+  const { nodes, setNodes, reactFlowWrapper } = useFlowStore();
  const { isPc } = useGlobalStore();
+  const { x, y, zoom } = useViewport();

  const filterTemplates = useMemo(() => {
    const guideModulesIndex = ModuleTemplates.findIndex((item) => item.label === '引导模块');
@@ -47,6 +43,28 @@ const ModuleTemplateList = ({
    ];
  }, [nodes]);

+  const onAddNode = useCallback(
+    ({ template, position }: { template: FlowModuleTemplateType; position: XYPosition }) => {
+      if (!reactFlowWrapper?.current) return;
+
+      const reactFlowBounds = reactFlowWrapper.current.getBoundingClientRect();
+      const mouseX = (position.x - reactFlowBounds.left - x) / zoom - 100;
+      const mouseY = (position.y - reactFlowBounds.top - y) / zoom;
+      setNodes((state) =>
+        state.concat(
+          appModule2FlowNode({
+            item: {
+              ...template,
+              moduleId: nanoid(),
+              position: { x: mouseX, y: mouseY }
+            }
+          })
+        )
+      );
+    },
+    [reactFlowWrapper, setNodes, x, y, zoom]
+  );
+
  return (
    <>
      <Box
--- a/projects/app/src/pages/app/detail/components/AdEdit/components/render/RenderInput.tsx
+++ b/projects/app/src/pages/app/detail/components/AdEdit/components/render/RenderInput.tsx
@@ -32,6 +32,7 @@ import { formatPrice } from '@fastgpt/common/bill';
 import { useDatasetStore } from '@/web/core/store/dataset';
 import { SelectedDatasetType } from '@/types/core/dataset';
 import { useQuery } from '@tanstack/react-query';
+import { LLMModelItemType } from '@/types/model';

 const SetInputFieldModal = dynamic(() => import('../modules/SetInputFieldModal'));
 const SelectAppModal = dynamic(() => import('../../../SelectAppModal'));
@@ -186,8 +187,8 @@ const RenderInput = ({
                {item.type === FlowInputItemTypeEnum.selectApp && (
                  <SelectAppRender item={item} moduleId={moduleId} />
                )}
-                {item.type === FlowInputItemTypeEnum.quoteList && (
-                  <QuoteListRender inputs={sortInputs} item={item} moduleId={moduleId} />
+                {item.type === FlowInputItemTypeEnum.aiSettings && (
+                  <AISetting inputs={sortInputs} item={item} moduleId={moduleId} />
                )}
                {item.type === FlowInputItemTypeEnum.maxToken && (
                  <MaxTokenRender inputs={sortInputs} item={item} moduleId={moduleId} />
@@ -343,7 +344,7 @@ var SliderRender = React.memo(function SliderRender({ item, moduleId }: RenderPr
  );
 });

-var QuoteListRender = React.memo(function QuoteListRender({ inputs = [], moduleId }: RenderProps) {
+var AISetting = React.memo(function AISetting({ inputs = [], moduleId }: RenderProps) {
  const { onChangeNode } = useFlowStore();
  const { t } = useTranslation();
  const chatModulesData = useMemo(() => {
@@ -367,10 +368,11 @@ var QuoteListRender = React.memo(function QuoteListRender({ inputs = [], moduleI
        leftIcon={<MyIcon name={'settingLight'} w={'14px'} />}
        onClick={onOpenAIChatSetting}
      >
-        {t('app.Quote Prompt Settings')}
+        {t('app.AI Settings')}
      </Button>
      {isOpenAIChatSetting && (
        <AIChatSettingsModal
+          isAdEdit
          onClose={onCloseAIChatSetting}
          onSuccess={(e) => {
            for (let key in e) {
@@ -404,7 +406,7 @@ var MaxTokenRender = React.memo(function MaxTokenRender({
  const { onChangeNode } = useFlowStore();
  const model = inputs.find((item) => item.key === 'model')?.value;
  const modelData = chatModelList.find((item) => item.model === model);
-  const maxToken = modelData ? modelData.contextMaxToken : 4000;
+  const maxToken = modelData ? modelData.maxToken : 4000;
  const markList = [
    { label: '100', value: 100 },
    { label: `${maxToken}`, value: maxToken }
@@ -441,8 +443,42 @@ var SelectChatModelRender = React.memo(function SelectChatModelRender({
  moduleId
 }: RenderProps) {
  const { onChangeNode } = useFlowStore();
+  const modelList = (item.customData?.() as LLMModelItemType[]) || chatModelList || [];

-  const list = chatModelList.map((item) => {
+  function onChangeModel(e: string) {
+    {
+      onChangeNode({
+        moduleId,
+        type: 'inputs',
+        key: item.key,
+        value: {
+          ...item,
+          value: e
+        }
+      });
+
+      // update max tokens
+      const model = modelList.find((item) => item.model === e) || modelList[0];
+      if (!model) return;
+
+      onChangeNode({
+        moduleId,
+        type: 'inputs',
+        key: 'maxToken',
+        value: {
+          ...inputs.find((input) => input.key === 'maxToken'),
+          markList: [
+            { label: '100', value: 100 },
+            { label: `${model.maxToken}`, value: model.maxToken }
+          ],
+          max: model.maxToken,
+          value: model.maxToken / 2
+        }
+      });
+    }
+  }
+
+  const list = modelList.map((item) => {
    const priceStr = `(${formatPrice(item.price, 1000)}元/1k Tokens)`;

    return {
@@ -451,43 +487,11 @@ var SelectChatModelRender = React.memo(function SelectChatModelRender({
    };
  });

-  return (
-    <MySelect
-      width={'100%'}
-      value={item.value}
-      list={list}
-      onchange={(e) => {
-        onChangeNode({
-          moduleId,
-          type: 'inputs',
-          key: item.key,
-          value: {
-            ...item,
-            value: e
-          }
-        });
+  if (!item.value && list.length > 0) {
+    onChangeModel(list[0].value);
+  }

-        // update max tokens
-        const model = chatModelList.find((item) => item.model === e) || chatModelList[0];
-        if (!model) return;
-
-        onChangeNode({
-          moduleId,
-          type: 'inputs',
-          key: 'maxToken',
-          value: {
-            ...inputs.find((input) => input.key === 'maxToken'),
-            markList: [
-              { label: '100', value: 100 },
-              { label: `${model.contextMaxToken}`, value: model.contextMaxToken }
-            ],
-            max: model.contextMaxToken,
-            value: model.contextMaxToken / 2
-          }
-        });
-      }}
-    />
-  );
+  return <MySelect width={'100%'} value={item.value} list={list} onchange={onChangeModel} />;
 });

 var SelectDatasetRender = React.memo(function SelectDatasetRender({ item, moduleId }: RenderProps) {
--- a/projects/app/src/pages/app/detail/components/AdEdit/index.tsx
+++ b/projects/app/src/pages/app/detail/components/AdEdit/index.tsx
@@ -25,6 +25,7 @@ import MyTooltip from '@/components/MyTooltip';
 import TemplateList from './components/TemplateList';
 import ChatTest, { type ChatTestComponentRef } from './components/ChatTest';
 import FlowProvider, { useFlowStore } from './components/Provider';
+import Header from './components/Header';

 const ImportSettings = dynamic(() => import('./components/ImportSettings'));
 const NodeChat = dynamic(() => import('./components/Nodes/NodeChat'));
@@ -62,187 +63,7 @@ const edgeTypes = {
 };
 type Props = { app: AppSchema; onCloseSettings: () => void };

-function FlowHeader({ app, onCloseSettings }: Props & {}) {
-  const theme = useTheme();
-  const { t } = useTranslation();
-  const { copyData } = useCopyData();
-  const ChatTestRef = useRef<ChatTestComponentRef>(null);
-  const { isOpen: isOpenImport, onOpen: onOpenImport, onClose: onCloseImport } = useDisclosure();
-  const { updateAppDetail } = useUserStore();
-  const { nodes, edges, onFixView } = useFlowStore();
-
-  const [testModules, setTestModules] = useState<AppModuleItemType[]>();
-
-  const flow2AppModules = useCallback(() => {
-    const modules: AppModuleItemType[] = nodes.map((item) => ({
-      moduleId: item.data.moduleId,
-      name: item.data.name,
-      flowType: item.data.flowType,
-      showStatus: item.data.showStatus,
-      position: item.position,
-      inputs: item.data.inputs.map((item) => ({
-        ...item,
-        connected: item.connected ?? item.type !== FlowInputItemTypeEnum.target
-      })),
-      outputs: item.data.outputs.map((item) => ({
-        ...item,
-        targets: [] as FlowOutputTargetItemType[]
-      }))
-    }));
-
-    // update inputs and outputs
-    modules.forEach((module) => {
-      module.inputs.forEach((input) => {
-        input.connected =
-          input.connected ||
-          !!edges.find(
-            (edge) => edge.target === module.moduleId && edge.targetHandle === input.key
-          );
-      });
-      module.outputs.forEach((output) => {
-        output.targets = edges
-          .filter(
-            (edge) =>
-              edge.source === module.moduleId &&
-              edge.sourceHandle === output.key &&
-              edge.targetHandle
-          )
-          .map((edge) => ({
-            moduleId: edge.target,
-            key: edge.targetHandle || ''
-          }));
-      });
-    });
-    return modules;
-  }, [edges, nodes]);
-
-  const { mutate: onclickSave, isLoading } = useRequest({
-    mutationFn: () => {
-      const modules = flow2AppModules();
-      // check required connect
-      for (let i = 0; i < modules.length; i++) {
-        const item = modules[i];
-        if (item.inputs.find((input) => input.required && !input.connected)) {
-          return Promise.reject(`【${item.name}】存在未连接的必填输入`);
-        }
-        if (item.inputs.find((input) => input.valueCheck && !input.valueCheck(input.value))) {
-          return Promise.reject(`【${item.name}】存在为填写的必填项`);
-        }
-      }
-
-      return updateAppDetail(app._id, {
-        modules: modules,
-        type: AppTypeEnum.advanced
-      });
-    },
-    successToast: '保存配置成功',
-    errorToast: '保存配置异常',
-    onSuccess() {
-      ChatTestRef.current?.resetChatTest();
-    }
-  });
-
-  return (
-    <>
-      <Flex
-        py={3}
-        px={[2, 5, 8]}
-        borderBottom={theme.borders.base}
-        alignItems={'center'}
-        userSelect={'none'}
-      >
-        <MyTooltip label={'返回'} offset={[10, 10]}>
-          <IconButton
-            size={'sm'}
-            icon={<MyIcon name={'back'} w={'14px'} />}
-            borderRadius={'md'}
-            borderColor={'myGray.300'}
-            variant={'base'}
-            aria-label={''}
-            onClick={() => {
-              onCloseSettings();
-              onFixView();
-            }}
-          />
-        </MyTooltip>
-        <Box ml={[3, 6]} fontSize={['md', '2xl']} flex={1}>
-          {app.name}
-        </Box>
-
-        <MyTooltip label={t('app.Import Configs')}>
-          <IconButton
-            mr={[3, 6]}
-            icon={<MyIcon name={'importLight'} w={['14px', '16px']} />}
-            borderRadius={'lg'}
-            variant={'base'}
-            aria-label={'save'}
-            onClick={onOpenImport}
-          />
-        </MyTooltip>
-        <MyTooltip label={t('app.Export Configs')}>
-          <IconButton
-            mr={[3, 6]}
-            icon={<MyIcon name={'export'} w={['14px', '16px']} />}
-            borderRadius={'lg'}
-            variant={'base'}
-            aria-label={'save'}
-            onClick={() =>
-              copyData(
-                JSON.stringify(flow2AppModules(), null, 2),
-                t('app.Export Config Successful')
-              )
-            }
-          />
-        </MyTooltip>
-
-        {testModules ? (
-          <IconButton
-            mr={[3, 6]}
-            icon={<SmallCloseIcon fontSize={'25px'} />}
-            variant={'base'}
-            color={'myGray.600'}
-            borderRadius={'lg'}
-            aria-label={''}
-            onClick={() => setTestModules(undefined)}
-          />
-        ) : (
-          <MyTooltip label={'测试对话'}>
-            <IconButton
-              mr={[3, 6]}
-              icon={<MyIcon name={'chat'} w={['14px', '16px']} />}
-              borderRadius={'lg'}
-              aria-label={'save'}
-              variant={'base'}
-              onClick={() => {
-                setTestModules(flow2AppModules());
-              }}
-            />
-          </MyTooltip>
-        )}
-
-        <MyTooltip label={'保存配置'}>
-          <IconButton
-            icon={<MyIcon name={'save'} w={['14px', '16px']} />}
-            borderRadius={'lg'}
-            isLoading={isLoading}
-            aria-label={'save'}
-            onClick={onclickSave}
-          />
-        </MyTooltip>
-      </Flex>
-      {isOpenImport && <ImportSettings onClose={onCloseImport} />}
-      <ChatTest
-        ref={ChatTestRef}
-        modules={testModules}
-        app={app}
-        onClose={() => setTestModules(undefined)}
-      />
-    </>
-  );
-}
-const Header = React.memo(FlowHeader);
-
-const AppEdit = (props: Props) => {
+const AppEdit = React.memo(function AppEdit(props: Props) {
  const { app } = props;

  const {
@@ -261,7 +82,7 @@ const AppEdit = (props: Props) => {
  return (
    <>
      {/* header */}
-      <Header {...props} />
+      <Header app={app} onCloseSettings={props.onCloseSettings} />
      <Box
        minH={'400px'}
        flex={'1 0 0'}
@@ -318,11 +139,11 @@ const AppEdit = (props: Props) => {
          <Controls position={'bottom-right'} style={{ display: 'flex' }} showInteractive={false} />
        </ReactFlow>

-        <TemplateList isOpen={isOpenTemplate} nodes={nodes} onClose={onCloseTemplate} />
+        <TemplateList isOpen={isOpenTemplate} onClose={onCloseTemplate} />
      </Box>
    </>
  );
-};
+});

 const Flow = (data: Props) => {
  return (
--- a/projects/app/src/pages/app/detail/components/BasicEdit/index.tsx
+++ b/projects/app/src/pages/app/detail/components/BasicEdit/index.tsx
@@ -34,7 +34,6 @@ import { chatModelList } from '@/web/common/store/static';
 import { formatPrice } from '@fastgpt/common/bill/index';
 import {
  ChatModelSystemTip,
-  ChatModelLimitTip,
  welcomeTextTip,
  questionGuideTip
 } from '@/constants/flow/ModuleTemplate';
@@ -128,12 +127,7 @@ const Settings = ({ appId }: { appId: string }) => {
      label: `${item.name} (${formatPrice(item.price, 1000)} 元/1k tokens)`
    }));
  }, [refresh]);
-  const tokenLimit = useMemo(() => {
-    return (
-      chatModelList.find((item) => item.model === getValues('chatModel.model'))?.contextMaxToken ||
-      4000
-    );
-  }, [getValues, refresh]);
+
  const selectedKbList = useMemo(
    () => allDatasets.filter((item) => kbList.find((kb) => kb.kbId === item._id)),
    [allDatasets, kbList]
@@ -411,6 +405,10 @@ const Settings = ({ appId }: { appId: string }) => {
          <Box ml={2} flex={1}>
            AI 配置
          </Box>
+          <Flex {...BoxBtnStyles} onClick={onOpenAIChatSetting}>
+            <MyIcon mr={1} name={'settingLight'} w={'14px'} />
+            高级配置
+          </Flex>
        </Flex>

        <Flex alignItems={'center'} mt={5}>
@@ -424,7 +422,7 @@ const Settings = ({ appId }: { appId: string }) => {
                setValue('chatModel.model', val);
                const maxToken =
                  chatModelList.find((item) => item.model === getValues('chatModel.model'))
-                    ?.contextMaxToken || 4000;
+                    ?.maxToken || 4000;
                const token = maxToken / 2;
                setValue('chatModel.maxToken', token);
                setRefresh(!refresh);
@@ -432,45 +430,6 @@ const Settings = ({ appId }: { appId: string }) => {
            />
          </Box>
        </Flex>
-        <Flex alignItems={'center'} my={10}>
-          <Box {...LabelStyles}>温度</Box>
-          <Box flex={1} ml={'10px'}>
-            <MySlider
-              markList={[
-                { label: '严谨', value: 0 },
-                { label: '发散', value: 10 }
-              ]}
-              width={'95%'}
-              min={0}
-              max={10}
-              value={getValues('chatModel.temperature')}
-              onChange={(e) => {
-                setValue('chatModel.temperature', e);
-                setRefresh(!refresh);
-              }}
-            />
-          </Box>
-        </Flex>
-        <Flex alignItems={'center'} mt={12} mb={10}>
-          <Box {...LabelStyles}>回复上限</Box>
-          <Box flex={1} ml={'10px'}>
-            <MySlider
-              markList={[
-                { label: '100', value: 100 },
-                { label: `${tokenLimit}`, value: tokenLimit }
-              ]}
-              width={'95%'}
-              min={100}
-              max={tokenLimit}
-              step={50}
-              value={getValues('chatModel.maxToken')}
-              onChange={(val) => {
-                setValue('chatModel.maxToken', val);
-                setRefresh(!refresh);
-              }}
-            />
-          </Box>
-        </Flex>
        <Flex mt={10} alignItems={'flex-start'}>
          <Box {...LabelStyles}>
            提示词
@@ -502,10 +461,6 @@ const Settings = ({ appId }: { appId: string }) => {
            <MyIcon name={'edit'} w={'14px'} mr={1} />
            参数
          </Flex>
-          <Flex {...BoxBtnStyles} onClick={onOpenAIChatSetting}>
-            <MyIcon mr={1} name={'settingLight'} w={'14px'} />
-            提示词
-          </Flex>
        </Flex>
        <Flex mt={1} color={'myGray.600'} fontSize={['sm', 'md']}>
          相似度: {getValues('kb.searchSimilarity')}, 单次搜索数量: {getValues('kb.searchLimit')},
--- a/projects/app/src/pages/kb/detail/components/Import/QA.tsx
+++ b/projects/app/src/pages/kb/detail/components/Import/QA.tsx
@@ -6,7 +6,7 @@ import { useMutation } from '@tanstack/react-query';
 import { splitText2Chunks } from '@/utils/file';
 import { getErrText } from '@/utils/tools';
 import { formatPrice } from '@fastgpt/common/bill/index';
-import { qaModel } from '@/web/common/store/static';
+import { qaModelList } from '@/web/common/store/static';
 import MyIcon from '@/components/Icon';
 import CloseIcon from '@/components/Icon/close';
 import DeleteIcon, { hoverDeleteStyles } from '@/components/Icon/delete';
@@ -23,8 +23,9 @@ import { chunksUpload } from '@/web/core/utils/dataset';
 const fileExtension = '.txt, .doc, .docx, .pdf, .md';

 const QAImport = ({ kbId }: { kbId: string }) => {
-  const unitPrice = qaModel.price || 3;
-  const chunkLen = qaModel.maxToken * 0.45;
+  const qaModel = qaModelList[0];
+  const unitPrice = qaModel?.price || 3;
+  const chunkLen = qaModel?.maxToken * 0.45;
  const theme = useTheme();
  const router = useRouter();
  const { toast } = useToast();
--- a/projects/app/src/pages/kb/list/component/CreateModal.tsx
+++ b/projects/app/src/pages/kb/list/component/CreateModal.tsx
@@ -13,9 +13,9 @@ import MyTooltip from '@/components/MyTooltip';
 import MyModal from '@/components/MyModal';
 import { postCreateDataset } from '@/web/core/api/dataset';
 import type { CreateDatasetParams } from '@/global/core/api/datasetReq.d';
-import { vectorModelList } from '@/web/common/store/static';
 import MySelect from '@/components/Select';
 import { QuestionOutlineIcon } from '@chakra-ui/icons';
+import { vectorModelList } from '@/web/common/store/static';
 import Tag from '@/components/Tag';

 const CreateModal = ({ onClose, parentId }: { onClose: () => void; parentId?: string }) => {
--- a/projects/app/src/service/common/bill/push.ts
+++ b/projects/app/src/service/common/bill/push.ts
@@ -1,12 +1,12 @@
 import { Bill } from '@/service/mongo';
 import { MongoUser } from '@fastgpt/support/user/schema';
 import { BillSourceEnum } from '@/constants/user';
-import { getModel } from '@/service/utils/data';
+import { getModelMap, ModelTypeEnum } from '@/service/core/ai/model';
 import { ChatHistoryItemResType } from '@/types/chat';
 import { formatPrice } from '@fastgpt/common/bill/index';
 import { addLog } from '@/service/utils/tools';
 import type { CreateBillType } from '@/types/common/bill';
-import { defaultQGModel } from '@/pages/api/system/getInitData';
+import { defaultQGModels } from '@/constants/model';

 async function createBill(data: CreateBillType) {
  try {
@@ -106,7 +106,7 @@ export const pushQABill = async ({
  addLog.info('splitData generate success', { totalTokens });

  // 获取模型单价格, 都是用 gpt35 拆分
-  const unitPrice = global.qaModel.price || 3;
+  const unitPrice = global.qaModels?.[0]?.price || 3;
  // 计算价格
  const total = unitPrice * totalTokens;

@@ -158,7 +158,7 @@ export const pushGenerateVectorBill = async ({
        {
          moduleName: '索引生成',
          amount: total,
-          model: vectorModel.model,
+          model: vectorModel.name,
          tokenLen
        }
      ]
@@ -167,14 +167,22 @@ export const pushGenerateVectorBill = async ({
  return { total };
 };

-export const countModelPrice = ({ model, tokens }: { model: string; tokens: number }) => {
-  const modelData = getModel(model);
+export const countModelPrice = ({
+  model,
+  tokens,
+  type
+}: {
+  model: string;
+  tokens: number;
+  type: `${ModelTypeEnum}`;
+}) => {
+  const modelData = getModelMap?.[type]?.(model);
  if (!modelData) return 0;
  return modelData.price * tokens;
 };

 export const pushQuestionGuideBill = ({ tokens, userId }: { tokens: number; userId: string }) => {
-  const qgModel = global.qgModel || defaultQGModel;
+  const qgModel = global.qgModels?.[0] || defaultQGModels[0];
  const total = qgModel.price * tokens;
  createBill({
    userId,
--- a/projects/app/src/service/core/ai/model.ts
+++ b/projects/app/src/service/core/ai/model.ts
@@ -0,0 +1,68 @@
+import {
+  defaultChatModels,
+  defaultCQModels,
+  defaultExtractModels,
+  defaultQAModels,
+  defaultQGModels,
+  defaultVectorModels
+} from '@/constants/model';
+
+export const getChatModel = (model?: string) => {
+  return (
+    (global.chatModels || defaultChatModels).find((item) => item.model === model) ||
+    defaultChatModels[0]
+  );
+};
+export const getQAModel = (model?: string) => {
+  return (
+    (global.qaModels || defaultQAModels).find((item) => item.model === model) ||
+    global.qaModels?.[0] ||
+    defaultQAModels[0]
+  );
+};
+export const getCQModel = (model?: string) => {
+  return (
+    (global.cqModels || defaultCQModels).find((item) => item.model === model) ||
+    global.cqModels?.[0] ||
+    defaultCQModels[0]
+  );
+};
+export const getExtractModel = (model?: string) => {
+  return (
+    (global.extractModels || defaultExtractModels).find((item) => item.model === model) ||
+    global.extractModels?.[0] ||
+    defaultExtractModels[0]
+  );
+};
+export const getQGModel = (model?: string) => {
+  return (
+    (global.qgModels || defaultQGModels).find((item) => item.model === model) ||
+    global.qgModels?.[0] ||
+    defaultQGModels[0]
+  );
+};
+
+export const getVectorModel = (model?: string) => {
+  return (
+    global.vectorModels.find((item) => item.model === model) ||
+    global.vectorModels?.[0] ||
+    defaultVectorModels[0]
+  );
+};
+
+export enum ModelTypeEnum {
+  chat = 'chat',
+  qa = 'qa',
+  cq = 'cq',
+  extract = 'extract',
+  qg = 'qg',
+  vector = 'vector'
+}
+export const getModelMap = {
+  [ModelTypeEnum.chat]: getChatModel,
+  [ModelTypeEnum.qa]: getQAModel,
+  [ModelTypeEnum.cq]: getCQModel,
+  [ModelTypeEnum.extract]: getExtractModel,
+  [ModelTypeEnum.qg]: getQGModel,
+  [ModelTypeEnum.vector]: getVectorModel
+};
--- a/projects/app/src/service/core/app/module.ts
+++ b/projects/app/src/service/core/app/module.ts
@@ -0,0 +1,12 @@
+import { FlowModuleTypeEnum } from '@/constants/flow';
+import { AppModuleItemType } from '@/types/app';
+
+export const getChatModelNameListByModules = (modules: AppModuleItemType[]): string[] => {
+  const chatModules = modules.filter((item) => item.flowType === FlowModuleTypeEnum.chatNode);
+  return chatModules
+    .map((item) => {
+      const model = item.inputs.find((input) => input.key === 'model')?.value;
+      return global.chatModels.find((item) => item.model === model)?.name || '';
+    })
+    .filter((item) => item);
+};
--- a/projects/app/src/service/events/generateQA.ts
+++ b/projects/app/src/service/events/generateQA.ts
@@ -73,7 +73,7 @@ export async function generateQA(): Promise<any> {
    ];
    const ai = getAIApi(undefined, 480000);
    const chatResponse = await ai.chat.completions.create({
-      model: global.qaModel.model,
+      model: global.qaModels[0].model,
      temperature: 0.01,
      messages,
      stream: false
--- a/projects/app/src/service/moduleDispatch/agent/classifyQuestion.ts
+++ b/projects/app/src/service/moduleDispatch/agent/classifyQuestion.ts
@@ -10,9 +10,11 @@ import { FlowModuleTypeEnum } from '@/constants/flow';
 import type { ModuleDispatchProps } from '@/types/core/chat/type';
 import { replaceVariable } from '@/utils/common/tools/text';
 import { Prompt_CQJson } from '@/global/core/prompt/agent';
-import { defaultCQModel } from '@/pages/api/system/getInitData';
+import { FunctionModelItemType } from '@/types/model';
+import { getCQModel } from '@/service/core/ai/model';

 type Props = ModuleDispatchProps<{
+  model: string;
  systemPrompt?: string;
  history?: ChatItemType[];
  [SystemInputEnum.userChatInput]: string;
@@ -30,20 +32,26 @@ export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse
  const {
    moduleName,
    user,
-    inputs: { agents, userChatInput }
+    inputs: { model, agents, userChatInput }
  } = props as Props;

  if (!userChatInput) {
    return Promise.reject('Input is empty');
  }

-  const cqModel = global.cqModel || defaultCQModel;
+  const cqModel = getCQModel(model);

  const { arg, tokens } = await (async () => {
    if (cqModel.functionCall) {
-      return functionCall(props);
+      return functionCall({
+        ...props,
+        cqModel
+      });
    }
-    return completions(props);
+    return completions({
+      ...props,
+      cqModel
+    });
  })();

  const result = agents.find((item) => item.key === arg?.type) || agents[agents.length - 1];
@@ -64,45 +72,45 @@ export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse

 async function functionCall({
  user,
+  cqModel,
  inputs: { agents, systemPrompt, history = [], userChatInput }
-}: Props) {
-  const cqModel = global.cqModel;
-
+}: Props & { cqModel: FunctionModelItemType }) {
  const messages: ChatItemType[] = [
-    ...(systemPrompt
-      ? [
-          {
-            obj: ChatRoleEnum.System,
-            value: systemPrompt
-          }
-        ]
-      : []),
    ...history,
    {
      obj: ChatRoleEnum.Human,
-      value: userChatInput
+      value: systemPrompt
+        ? `补充的背景知识:
+"""
+${systemPrompt}
+"""
+我的问题: ${userChatInput}
+      `
+        : userChatInput
    }
  ];
+
  const filterMessages = ChatContextFilter({
    messages,
    maxTokens: cqModel.maxToken
  });
  const adaptMessages = adaptChat2GptMessages({ messages: filterMessages, reserveId: false });

-  //   function body
+  // function body
  const agentFunction = {
    name: agentFunName,
-    description: '判断用户问题的类型属于哪方面，返回对应的字段',
+    description: '请根据对话记录及补充的背景知识，判断用户的问题类型，并返回对应的字段',
    parameters: {
      type: 'object',
      properties: {
        type: {
          type: 'string',
-          description: agents.map((item) => `${item.value}，返回：'${item.key}'`).join('；'),
+          description: `判断用户的问题类型，并返回对应的字段。下面是几种问题类型: ${agents
+            .map((item) => `${item.value}，返回：'${item.key}'`)
+            .join('；')}`,
          enum: agents.map((item) => item.key)
        }
-      },
-      required: ['type']
+      }
    }
  };
  const ai = getAIApi(user.openaiAccount, 48000);
@@ -133,15 +141,14 @@ async function functionCall({
 }

 async function completions({
+  cqModel,
  user,
  inputs: { agents, systemPrompt = '', history = [], userChatInput }
-}: Props) {
-  const extractModel = global.extractModel;
-
+}: Props & { cqModel: FunctionModelItemType }) {
  const messages: ChatItemType[] = [
    {
      obj: ChatRoleEnum.Human,
-      value: replaceVariable(extractModel.prompt || Prompt_CQJson, {
+      value: replaceVariable(cqModel.functionPrompt || Prompt_CQJson, {
        systemPrompt,
        typeList: agents.map((item) => `ID: "${item.key}", 问题类型:${item.value}`).join('\n'),
        text: `${history.map((item) => `${item.obj}:${item.value}`).join('\n')}
@@ -153,7 +160,7 @@ Human:${userChatInput}`
  const ai = getAIApi(user.openaiAccount, 480000);

  const data = await ai.chat.completions.create({
-    model: extractModel.model,
+    model: cqModel.model,
    temperature: 0.01,
    messages: adaptChat2GptMessages({ messages, reserveId: false }),
    stream: false
--- a/projects/app/src/service/moduleDispatch/agent/extract.ts
+++ b/projects/app/src/service/moduleDispatch/agent/extract.ts
@@ -9,7 +9,7 @@ import { FlowModuleTypeEnum } from '@/constants/flow';
 import type { ModuleDispatchProps } from '@/types/core/chat/type';
 import { Prompt_ExtractJson } from '@/global/core/prompt/agent';
 import { replaceVariable } from '@/utils/common/tools/text';
-import { defaultExtractModel } from '@/pages/api/system/getInitData';
+import { FunctionModelItemType } from '@/types/model';

 type Props = ModuleDispatchProps<{
  history?: ChatItemType[];
@@ -37,13 +37,19 @@ export async function dispatchContentExtract(props: Props): Promise<Response> {
    return Promise.reject('Input is empty');
  }

-  const extractModel = global.extractModel || defaultExtractModel;
+  const extractModel = global.extractModels[0];

  const { arg, tokens } = await (async () => {
    if (extractModel.functionCall) {
-      return functionCall(props);
+      return functionCall({
+        ...props,
+        extractModel
+      });
    }
-    return completions(props);
+    return completions({
+      ...props,
+      extractModel
+    });
  })();

  // remove invalid key
@@ -83,11 +89,10 @@ export async function dispatchContentExtract(props: Props): Promise<Response> {
 }

 async function functionCall({
+  extractModel,
  user,
  inputs: { history = [], content, extractKeys, description }
-}: Props) {
-  const extractModel = global.extractModel;
-
+}: Props & { extractModel: FunctionModelItemType }) {
  const messages: ChatItemType[] = [
    ...history,
    {
@@ -152,15 +157,14 @@ async function functionCall({
 }

 async function completions({
+  extractModel,
  user,
  inputs: { history = [], content, extractKeys, description }
-}: Props) {
-  const extractModel = global.extractModel;
-
+}: Props & { extractModel: FunctionModelItemType }) {
  const messages: ChatItemType[] = [
    {
      obj: ChatRoleEnum.Human,
-      value: replaceVariable(extractModel.prompt || Prompt_ExtractJson, {
+      value: replaceVariable(extractModel.functionPrompt || Prompt_ExtractJson, {
        description,
        json: extractKeys
          .map(
--- a/projects/app/src/service/moduleDispatch/chat/oneapi.ts
+++ b/projects/app/src/service/moduleDispatch/chat/oneapi.ts
@@ -7,7 +7,6 @@ import { textAdaptGptResponse } from '@/utils/adapt';
 import { getAIApi } from '@fastgpt/core/ai/config';
 import type { ChatCompletion, StreamChatType } from '@fastgpt/core/ai/type';
 import { TaskResponseKeyEnum } from '@/constants/chat';
-import { getChatModel } from '@/service/utils/data';
 import { countModelPrice } from '@/service/common/bill/push';
 import { ChatModelItemType } from '@/types/model';
 import { postTextCensor } from '@fastgpt/common/plusApi/censor';
@@ -15,12 +14,13 @@ import { ChatCompletionRequestMessageRoleEnum } from '@fastgpt/core/ai/constant'
 import { AppModuleItemType } from '@/types/app';
 import { countMessagesTokens, sliceMessagesTB } from '@/utils/common/tiktoken';
 import { adaptChat2GptMessages } from '@/utils/common/adapt/message';
-import { defaultQuotePrompt, defaultQuoteTemplate } from '@/global/core/prompt/AIChat';
+import { Prompt_QuotePromptList, Prompt_QuoteTemplateList } from '@/global/core/prompt/AIChat';
 import type { AIChatProps } from '@/types/core/aiChat';
 import { replaceVariable } from '@/utils/common/tools/text';
 import { FlowModuleTypeEnum } from '@/constants/flow';
 import type { ModuleDispatchProps } from '@/types/core/chat/type';
-import { responseWrite, responseWriteController } from '@/service/common/stream';
+import { responseWrite, responseWriteController } from '@fastgpt/common/tools/stream';
+import { getChatModel, ModelTypeEnum } from '@/service/core/ai/model';

 export type ChatProps = ModuleDispatchProps<
  AIChatProps & {
@@ -47,12 +47,13 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
    user,
    outputs,
    inputs: {
-      model = global.chatModels[0]?.model,
+      model,
      temperature = 0,
      maxToken = 4000,
      history = [],
      quoteQA = [],
      userChatInput,
+      isResponseAnswerText = true,
      systemPrompt = '',
      limitPrompt,
      quoteTemplate,
@@ -63,6 +64,8 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
    return Promise.reject('Question is empty');
  }

+  stream = stream && isResponseAnswerText;
+
  // temperature adapt
  const modelConstantsData = getChatModel(model);

@@ -110,18 +113,18 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
    model,
    temperature,
    max_tokens,
+    stream,
    messages: [
-      ...(modelConstantsData.defaultSystem
+      ...(modelConstantsData.defaultSystemChatPrompt
        ? [
            {
              role: ChatCompletionRequestMessageRoleEnum.System,
-              content: modelConstantsData.defaultSystem
+              content: modelConstantsData.defaultSystemChatPrompt
            }
          ]
        : []),
      ...messages
-    ],
-    stream
+    ]
  });

  const { answerText, totalTokens, completeMessages } = await (async () => {
@@ -172,7 +175,9 @@ export const dispatchChatCompletion = async (props: ChatProps): Promise<ChatResp
    [TaskResponseKeyEnum.responseData]: {
      moduleType: FlowModuleTypeEnum.chatNode,
      moduleName,
-      price: user.openaiAccount?.key ? 0 : countModelPrice({ model, tokens: totalTokens }),
+      price: user.openaiAccount?.key
+        ? 0
+        : countModelPrice({ model, tokens: totalTokens, type: ModelTypeEnum.chat }),
      model: modelConstantsData.name,
      tokens: totalTokens,
      question: userChatInput,
@@ -198,7 +203,7 @@ function filterQuote({
    maxTokens: model.quoteMaxToken,
    messages: quoteQA.map((item, index) => ({
      obj: ChatRoleEnum.System,
-      value: replaceVariable(quoteTemplate || defaultQuoteTemplate, {
+      value: replaceVariable(quoteTemplate || Prompt_QuoteTemplateList[0].value, {
        ...item,
        index: index + 1
      })
@@ -212,7 +217,7 @@ function filterQuote({
    filterQuoteQA.length > 0
      ? `${filterQuoteQA
          .map((item, index) =>
-            replaceVariable(quoteTemplate || defaultQuoteTemplate, {
+            replaceVariable(quoteTemplate || Prompt_QuoteTemplateList[0].value, {
              ...item,
              index: `${index + 1}`
            })
@@ -243,7 +248,7 @@ function getChatMessages({
  model: ChatModelItemType;
 }) {
  const question = quoteText
-    ? replaceVariable(quotePrompt || defaultQuotePrompt, {
+    ? replaceVariable(quotePrompt || Prompt_QuotePromptList[0].value, {
        quote: quoteText,
        question: userChatInput
      })
@@ -275,7 +280,7 @@ function getChatMessages({

  const filterMessages = ChatContextFilter({
    messages,
-    maxTokens: Math.ceil(model.contextMaxToken - 300) // filter token. not response maxToken
+    maxTokens: Math.ceil(model.maxToken - 300) // filter token. not response maxToken
  });

  const adaptMessages = adaptChat2GptMessages({ messages: filterMessages, reserveId: false });
@@ -294,7 +299,7 @@ function getMaxTokens({
  model: ChatModelItemType;
  filterMessages: ChatProps['inputs']['history'];
 }) {
-  const tokensLimit = model.contextMaxToken;
+  const tokensLimit = model.maxToken;
  /* count response max token */

  const promptsToken = countMessagesTokens({
@@ -349,7 +354,7 @@ async function streamResponse({
      stream.controller?.abort();
      break;
    }
-    const content = part.choices[0]?.delta?.content || '';
+    const content = part.choices?.[0]?.delta?.content || '';
    answer += content;

    responseWrite({
--- a/projects/app/src/service/moduleDispatch/kb/search.ts
+++ b/projects/app/src/service/moduleDispatch/kb/search.ts
@@ -8,6 +8,7 @@ import type { QuoteItemType } from '@/types/chat';
 import { PgDatasetTableName } from '@/constants/plugin';
 import { FlowModuleTypeEnum } from '@/constants/flow';
 import type { ModuleDispatchProps } from '@/types/core/chat/type';
+import { ModelTypeEnum } from '@/service/core/ai/model';
 type KBSearchProps = ModuleDispatchProps<{
  kbList: SelectedDatasetType;
  similarity: number;
@@ -66,7 +67,11 @@ export async function dispatchKBSearch(props: Record<string, any>): Promise<KBSe
    responseData: {
      moduleType: FlowModuleTypeEnum.kbSearchNode,
      moduleName,
-      price: countModelPrice({ model: vectorModel.model, tokens: tokenLen }),
+      price: countModelPrice({
+        model: vectorModel.model,
+        tokens: tokenLen,
+        type: ModelTypeEnum.vector
+      }),
      model: vectorModel.name,
      tokens: tokenLen,
      similarity,
--- a/projects/app/src/service/moduleDispatch/tools/answer.ts
+++ b/projects/app/src/service/moduleDispatch/tools/answer.ts
@@ -1,5 +1,5 @@
 import { sseResponseEventEnum, TaskResponseKeyEnum } from '@/constants/chat';
-import { sseResponse } from '@/service/utils/tools';
+import { responseWrite } from '@fastgpt/common/tools/stream';
 import { textAdaptGptResponse } from '@/utils/adapt';
 import type { ModuleDispatchProps } from '@/types/core/chat/type';
 export type AnswerProps = ModuleDispatchProps<{
@@ -21,7 +21,7 @@ export const dispatchAnswer = (props: Record<string, any>): AnswerResponse => {
  const formatText = typeof text === 'string' ? text : JSON.stringify(text, null, 2);

  if (stream) {
-    sseResponse({
+    responseWrite({
      res,
      event: detail ? sseResponseEventEnum.answer : undefined,
      data: textAdaptGptResponse({
--- a/projects/app/src/service/moduleDispatch/tools/runApp.ts
+++ b/projects/app/src/service/moduleDispatch/tools/runApp.ts
@@ -3,7 +3,7 @@ import type { ModuleDispatchProps } from '@/types/core/chat/type';
 import { SelectAppItemType } from '@/types/core/app/flow';
 import { dispatchModules } from '@/pages/api/v1/chat/completions';
 import { App } from '@/service/mongo';
-import { responseWrite } from '@/service/common/stream';
+import { responseWrite } from '@fastgpt/common/tools/stream';
 import { ChatRoleEnum, TaskResponseKeyEnum, sseResponseEventEnum } from '@/constants/chat';
 import { textAdaptGptResponse } from '@/utils/adapt';

--- a/projects/app/src/service/pg.ts
+++ b/projects/app/src/service/pg.ts
@@ -232,6 +232,6 @@ export async function initPg() {
    `);
    console.log('init pg successful');
  } catch (error) {
-    addLog.error('init pg error', error);
+    console.log('init pg error', error);
  }
 }
--- a/projects/app/src/service/response.ts
+++ b/projects/app/src/service/response.ts
@@ -1,7 +1,9 @@
 import { sseResponseEventEnum } from '@/constants/chat';
 import { NextApiResponse } from 'next';
 import { proxyError, ERROR_RESPONSE, ERROR_ENUM } from '@fastgpt/common/constant/errorCode';
-import { clearCookie, sseResponse, addLog } from './utils/tools';
+import { addLog } from './utils/tools';
+import { clearCookie } from '@fastgpt/support/user/auth';
+import { responseWrite } from '@fastgpt/common/tools/stream';

 export interface ResponseType<T = any> {
  code: number;
@@ -66,7 +68,7 @@ export const sseErrRes = (res: NextApiResponse, error: any) => {
      clearCookie(res);
    }

-    return sseResponse({
+    return responseWrite({
      res,
      event: sseResponseEventEnum.error,
      data: JSON.stringify(ERROR_RESPONSE[errResponseKey])
@@ -86,7 +88,7 @@ export const sseErrRes = (res: NextApiResponse, error: any) => {

  addLog.error(`sse error: ${msg}`, error);

-  sseResponse({
+  responseWrite({
    res,
    event: sseResponseEventEnum.error,
    data: JSON.stringify({ message: msg })
--- a/projects/app/src/service/utils/data.ts
+++ b/projects/app/src/service/utils/data.ts
@@ -1,24 +0,0 @@
-export const getChatModel = (model?: string) => {
-  return global.chatModels.find((item) => item.model === model);
-};
-export const getVectorModel = (model?: string) => {
-  return (
-    global.vectorModels.find((item) => item.model === model) || {
-      model: 'UnKnow',
-      name: 'UnKnow',
-      defaultToken: 500,
-      price: 0,
-      maxToken: 3000
-    }
-  );
-};
-
-export const getModel = (model?: string) => {
-  return [
-    ...global.chatModels,
-    ...global.vectorModels,
-    global.qaModel,
-    global.extractModel,
-    global.cqModel
-  ].find((item) => item.model === model);
-};
--- a/projects/app/src/service/utils/tools.ts
+++ b/projects/app/src/service/utils/tools.ts
@@ -1,37 +1,7 @@
-import type { NextApiResponse, NextApiHandler, NextApiRequest } from 'next';
-import NextCors from 'nextjs-cors';
+import type { NextApiResponse } from 'next';
 import { generateQA } from '../events/generateQA';
 import { generateVector } from '../events/generateVector';

-/* set cookie */
-export const setCookie = (res: NextApiResponse, token: string) => {
-  res.setHeader(
-    'Set-Cookie',
-    `token=${token}; Path=/; HttpOnly; Max-Age=604800; Samesite=None; Secure;`
-  );
-};
-/* clear cookie */
-export const clearCookie = (res: NextApiResponse) => {
-  res.setHeader('Set-Cookie', 'token=; Path=/; Max-Age=0');
-};
-
-export function withNextCors(handler: NextApiHandler): NextApiHandler {
-  return async function nextApiHandlerWrappedWithNextCors(
-    req: NextApiRequest,
-    res: NextApiResponse
-  ) {
-    const methods = ['GET', 'eHEAD', 'PUT', 'PATCH', 'POST', 'DELETE'];
-    const origin = req.headers.origin;
-    await NextCors(req, res, {
-      methods,
-      origin: origin,
-      optionsSuccessStatus: 200
-    });
-
-    return handler(req, res);
-  };
-}
-
 /* start task */
 export const startQueue = () => {
  if (!global.systemEnv) return;
@@ -43,20 +13,6 @@ export const startQueue = () => {
  }
 };

-export const sseResponse = ({
-  res,
-  event,
-  data
-}: {
-  res: NextApiResponse;
-  event?: string;
-  data: string;
-}) => {
-  if (res.closed) return;
-  event && res.write(`event: ${event}\n`);
-  res.write(`data: ${data}\n\n`);
-};
-
 /* add logger */
 export const addLog = {
  info: (msg: string, obj?: Record<string, any>) => {
--- a/projects/app/src/types/core/aiChat.d.ts
+++ b/projects/app/src/types/core/aiChat.d.ts
@@ -1,9 +1,12 @@
+import { SystemInputEnum } from '@/constants/app';
+
 /* ai chat modules props */
 export type AIChatProps = {
  model: string;
  systemPrompt?: string;
  temperature: number;
  maxToken: number;
+  [SystemInputEnum.isResponseAnswerText]: boolean;
  quoteTemplate?: string;
  quotePrompt?: string;
  frequency: number;
--- a/projects/app/src/types/core/app/flow.d.ts
+++ b/projects/app/src/types/core/app/flow.d.ts
@@ -31,6 +31,7 @@ export type FlowInputItemType = {
  required?: boolean;
  list?: { label: string; value: any }[];
  markList?: { label: string; value: any }[];
+  customData?: () => any;
  valueCheck?: (value: any) => boolean;
 };

--- a/projects/app/src/types/index.d.ts
+++ b/projects/app/src/types/index.d.ts
@@ -3,7 +3,7 @@ import type { Tiktoken } from 'js-tiktoken';
 import {
  ChatModelItemType,
  FunctionModelItemType,
-  QAModelItemType,
+  LLMModelItemType,
  VectorModelItemType
 } from './model';
 import { TrackEventName } from '@/constants/common';
@@ -36,10 +36,10 @@ declare global {

  var vectorModels: VectorModelItemType[];
  var chatModels: ChatModelItemType[];
-  var qaModel: QAModelItemType;
-  var extractModel: FunctionModelItemType;
-  var cqModel: FunctionModelItemType;
-  var qgModel: FunctionModelItemType;
+  var qaModels: LLMModelItemType[];
+  var cqModels: FunctionModelItemType[];
+  var extractModels: FunctionModelItemType[];
+  var qgModels: LLMModelItemType[];

  var priceMd: string;
  var systemVersion: string;
--- a/projects/app/src/types/model.d.ts
+++ b/projects/app/src/types/model.d.ts
@@ -1,19 +1,23 @@
-export type ChatModelItemType = {
-  model: string;
-  name: string;
-  contextMaxToken: number;
-  quoteMaxToken: number;
-  maxTemperature: number;
-  price: number;
-  censor?: boolean;
-  defaultSystem?: string;
-};
-export type QAModelItemType = {
+import { LLMModelUsageEnum } from '@/constants/model';
+
+export type LLMModelItemType = {
  model: string;
  name: string;
  maxToken: number;
  price: number;
 };
+export type ChatModelItemType = LLMModelItemType & {
+  quoteMaxToken: number;
+  maxTemperature: number;
+  censor?: boolean;
+  defaultSystemChatPrompt?: string;
+};
+
+export type FunctionModelItemType = LLMModelItemType & {
+  functionCall: boolean;
+  functionPrompt: string;
+};
+
 export type VectorModelItemType = {
  model: string;
  name: string;
@@ -21,11 +25,3 @@ export type VectorModelItemType = {
  price: number;
  maxToken: number;
 };
-export type FunctionModelItemType = {
-  model: string;
-  name: string;
-  maxToken: number;
-  price: number;
-  prompt: string;
-  functionCall: boolean;
-};
--- a/projects/app/src/utils/app.ts
+++ b/projects/app/src/utils/app.ts
@@ -36,9 +36,10 @@ export const getDefaultAppForm = (): EditFormType => {
      model: defaultChatModel?.model,
      systemPrompt: '',
      temperature: 0,
+      [SystemInputEnum.isResponseAnswerText]: true,
      quotePrompt: '',
      quoteTemplate: '',
-      maxToken: defaultChatModel ? defaultChatModel.contextMaxToken / 2 : 4000,
+      maxToken: defaultChatModel ? defaultChatModel.maxToken / 2 : 4000,
      frequency: 0.5,
      presence: -0.5
    },
@@ -185,6 +186,13 @@ const chatModelInput = (formData: EditFormType): FlowInputItemType[] => [
    label: '系统提示词',
    connected: true
  },
+  {
+    key: SystemInputEnum.isResponseAnswerText,
+    value: true,
+    type: 'hidden',
+    label: '返回AI内容',
+    connected: true
+  },
  {
    key: 'quoteTemplate',
    value: formData.chatModel.quoteTemplate || '',
@@ -328,7 +336,7 @@ const simpleChatTemplate = (formData: EditFormType): AppModuleItemType[] => [
    outputs: [
      {
        key: 'answerText',
-        label: '模型回复',
+        label: 'AI回复',
        description: '直接响应，无需配置',
        type: 'hidden',
        targets: []
@@ -533,7 +541,7 @@ const kbTemplate = (formData: EditFormType): AppModuleItemType[] => [
    outputs: [
      {
        key: 'answerText',
-        label: '模型回复',
+        label: 'AI回复',
        description: '直接响应，无需配置',
        type: 'hidden',
        targets: []
--- a/projects/app/src/utils/file.ts
+++ b/projects/app/src/utils/file.ts
@@ -12,11 +12,12 @@ export const splitText2Chunks = ({ text = '', maxLen }: { text: string; maxLen:
  const tempMarker = 'SPLIT_HERE_SPLIT_HERE';

  const stepReg: Record<number, RegExp> = {
-    0: /(\n)/g,
-    1: /([。]|\.\s)/g,
-    2: /([！？]|!\s|\?\s)/g,
-    3: /([；]|;\s)/g,
-    4: /([，]|,\s)/g
+    0: /(\n\n)/g,
+    1: /([\n])/g,
+    2: /([。]|\.\s)/g,
+    3: /([！？]|!\s|\?\s)/g,
+    4: /([；]|;\s)/g,
+    5: /([，]|,\s)/g
  };

  const splitTextRecursively = ({ text = '', step }: { text: string; step: number }) => {
@@ -43,7 +44,6 @@ export const splitText2Chunks = ({ text = '', maxLen }: { text: string; maxLen:
      .filter((part) => part);

    let chunks: string[] = [];
-
    let preChunk = '';
    let chunk = '';
    for (let i = 0; i < splitTexts.length; i++) {
--- a/projects/app/src/web/common/store/static.ts
+++ b/projects/app/src/web/common/store/static.ts
@@ -1,34 +1,41 @@
-import {
-  type QAModelItemType,
-  type ChatModelItemType,
-  type VectorModelItemType
-} from '@/types/model';
 import type { InitDateResponse } from '@/global/common/api/systemRes';
 import { getSystemInitData } from '@/web/common/api/system';
 import { delay } from '@/utils/tools';
 import type { FeConfigsType } from '@fastgpt/common/type/index.d';
+import {
+  defaultChatModels,
+  defaultQAModels,
+  defaultCQModels,
+  defaultExtractModels,
+  defaultQGModels,
+  defaultVectorModels
+} from '@/constants/model';

-export let chatModelList: ChatModelItemType[] = [];
-export let qaModel: QAModelItemType = {
-  model: 'gpt-3.5-turbo-16k',
-  name: 'GPT35-16k',
-  maxToken: 16000,
-  price: 0
-};
-export let vectorModelList: VectorModelItemType[] = [];
 export let feConfigs: FeConfigsType = {};
 export let priceMd = '';
 export let systemVersion = '0.0.0';

+export let vectorModelList = defaultVectorModels;
+export let chatModelList = defaultChatModels;
+export let qaModelList = defaultQAModels;
+export let cqModelList = defaultCQModels;
+export let extractModelList = defaultExtractModels;
+export let qgModelList = defaultQGModels;
+
 let retryTimes = 3;

 export const clientInitData = async (): Promise<InitDateResponse> => {
  try {
    const res = await getSystemInitData();

-    chatModelList = res.chatModels;
-    qaModel = res.qaModel;
-    vectorModelList = res.vectorModels;
+    chatModelList = res.chatModels || [];
+    qaModelList = res.qaModels || [];
+    cqModelList = res.cqModels || [];
+    extractModelList = res.extractModels || [];
+    qgModelList = res.qgModels || [];
+
+    vectorModelList = res.vectorModels || [];
+
    feConfigs = res.feConfigs;
    priceMd = res.priceMd;
    systemVersion = res.systemVersion;