diff --git a/admin/service/route/app.js b/admin/service/route/app.js
index aa088c77f..ba5162ff8 100644
--- a/admin/service/route/app.js
+++ b/admin/service/route/app.js
@@ -39,7 +39,6 @@ export const useAppRoute = (app) => {
           userId: app.userId,
           name: app.name,
           intro: app.intro,
-          app: app.chat?.chatModel,
           relatedKbs: kbNames, // 将relatedKbs的id转换为相应的Kb名称
           systemPrompt: app.chat?.systemPrompt || '',
           temperature: app.chat?.temperature || 0,
diff --git a/admin/service/schema.js b/admin/service/schema.js
index 7d427d8f5..8557dcae3 100644
--- a/admin/service/schema.js
+++ b/admin/service/schema.js
@@ -62,12 +62,6 @@ const appSchema = new mongoose.Schema({
   avatar: String,
   status: String,
   intro: String,
-  chat: {
-    relatedKbs: [mongoose.Schema.Types.ObjectId],
-    systemPrompt: String,
-    temperature: Number,
-    chatModel: String
-  },
   share: {
     topNum: Number,
     isShare: Boolean,
diff --git a/client/data/ChatModels.json b/client/data/ChatModels.json
new file mode 100644
index 000000000..d8aa6b8a7
--- /dev/null
+++ b/client/data/ChatModels.json
@@ -0,0 +1,26 @@
+{
+  "Gpt35-4k": {
+    "model": "gpt-3.5-turbo",
+    "name": "Gpt35-4k",
+    "contextMaxToken": 4000,
+    "systemMaxToken": 2400,
+    "maxTemperature": 1.2,
+    "price": 1.5
+  },
+  "Gpt35-16k": {
+    "model": "gpt-3.5-turbo-16k",
+    "name": "Gpt35-16k",
+    "contextMaxToken": 16000,
+    "systemMaxToken": 8000,
+    "maxTemperature": 1.2,
+    "price": 3
+  },
+  "Gpt4": {
+    "model": "gpt-4",
+    "name": "Gpt4",
+    "contextMaxToken": 8000,
+    "systemMaxToken": 4000,
+    "maxTemperature": 1.2,
+    "price": 45
+  }
+}
diff --git a/client/data/QAModels.json b/client/data/QAModels.json
new file mode 100644
index 000000000..42e96848f
--- /dev/null
+++ b/client/data/QAModels.json
@@ -0,0 +1,8 @@
+{
+  "Gpt35-16k": {
+    "model": "gpt-3.5-turbo-16k",
+    "name": "Gpt35-16k",
+    "maxToken": 16000,
+    "price": 3
+  }
+}
diff --git a/client/data/SystemParams.json b/client/data/SystemParams.json
new file mode 100644
index 000000000..08a807e07
--- /dev/null
+++ b/client/data/SystemParams.json
@@ -0,0 +1,6 @@
+{
+  "vectorMaxProcess": 10,
+  "qaMaxProcess": 10,
+  "pgIvfflatProbe": 10,
+  "sensitiveCheck": false
+}
diff --git a/client/data/VectorModels.json b/client/data/VectorModels.json
new file mode 100644
index 000000000..3e3aee8b0
--- /dev/null
+++ b/client/data/VectorModels.json
@@ -0,0 +1,7 @@
+{
+  "text-embedding-ada-002": {
+    "model": "text-embedding-ada-002",
+    "name": "Embedding-2",
+    "price": 0.2
+  }
+}
diff --git a/client/public/js/baidutongji.js b/client/public/js/baidutongji.js
deleted file mode 100644
index 632c5ee9c..000000000
--- a/client/public/js/baidutongji.js
+++ /dev/null
@@ -1,8 +0,0 @@
-var _hmt = _hmt || [];
-
-(function () {
-  const hm = document.createElement('script');
-  hm.src = 'https://hm.baidu.com/hm.js?a5357e9dab086658bac0b6faf148882e';
-  const s = document.getElementsByTagName('script')[0];
-  s.parentNode.insertBefore(hm, s);
-})();
diff --git a/client/src/api/plugins/kb.ts b/client/src/api/plugins/kb.ts
index f01efdb7b..9dd391d5e 100644
--- a/client/src/api/plugins/kb.ts
+++ b/client/src/api/plugins/kb.ts
@@ -2,7 +2,6 @@ import { GET, POST, PUT, DELETE } from '../request';
 import type { KbItemType } from '@/types/plugin';
 import { RequestPaging } from '@/types/index';
 import { TrainingModeEnum } from '@/constants/plugin';
-import { type QuoteItemType } from '@/pages/api/openapi/kb/appKbSearch';
 import {
   Props as PushDataProps,
   Response as PushDateResponse
@@ -60,7 +59,7 @@ export const getTrainingData = (data: { kbId: string; init: boolean }) =>
   }>(`/plugins/kb/data/getTrainingData`, data);
 
 export const getKbDataItemById = (dataId: string) =>
-  GET<QuoteItemType>(`/plugins/kb/data/getDataById`, { dataId });
+  GET(`/plugins/kb/data/getDataById`, { dataId });
 
 /**
  * 直接push数据
diff --git a/client/src/api/system.ts b/client/src/api/system.ts
index e46006474..cf384ace5 100644
--- a/client/src/api/system.ts
+++ b/client/src/api/system.ts
@@ -1,9 +1,6 @@
 import { GET, POST, PUT } from './request';
-import type { ChatModelItemType } from '@/constants/model';
 import type { InitDateResponse } from '@/pages/api/system/getInitData';
 
 export const getInitData = () => GET<InitDateResponse>('/system/getInitData');
 
-export const getSystemModelList = () => GET<ChatModelItemType[]>('/system/getModels');
-
 export const uploadImg = (base64Img: string) => POST<string>('/system/uploadImage', { base64Img });
diff --git a/client/src/components/ChatBox/index.tsx b/client/src/components/ChatBox/index.tsx
index b7aedcbd3..5946ce462 100644
--- a/client/src/components/ChatBox/index.tsx
+++ b/client/src/components/ChatBox/index.tsx
@@ -10,7 +10,7 @@ import React, {
 import { throttle } from 'lodash';
 import { ChatItemType, ChatSiteItemType, ExportChatType } from '@/types/chat';
 import { useToast } from '@/hooks/useToast';
-import { useCopyData, voiceBroadcast, hasVoiceApi } from '@/utils/tools';
+import { useCopyData, voiceBroadcast, hasVoiceApi, getErrText } from '@/utils/tools';
 import { Box, Card, Flex, Input, Textarea, Button, useTheme } from '@chakra-ui/react';
 import { useUserStore } from '@/store/user';
 
@@ -241,33 +241,34 @@ const ChatBox = (
           variables: data
         });
 
-        // 设置聊天内容为完成状态
-        setChatHistory((state) =>
-          state.map((item, index) => {
-            if (index !== state.length - 1) return item;
-            return {
-              ...item,
-              status: 'finish'
-            };
-          })
-        );
-
         setTimeout(() => {
           generatingScroll();
           TextareaDom.current?.focus();
         }, 100);
       } catch (err: any) {
         toast({
-          title: typeof err === 'string' ? err : err?.message || '聊天出错了~',
-          status: 'warning',
+          title: getErrText(err, '聊天出错了~'),
+          status: 'error',
           duration: 5000,
           isClosable: true
         });
 
-        resetInputVal(value);
-
-        setChatHistory(newChatList.slice(0, newChatList.length - 2));
+        if (!err?.responseText) {
+          resetInputVal(value);
+          setChatHistory(newChatList.slice(0, newChatList.length - 2));
+        }
       }
+
+      // set finish status
+      setChatHistory((state) =>
+        state.map((item, index) => {
+          if (index !== state.length - 1) return item;
+          return {
+            ...item,
+            status: 'finish'
+          };
+        })
+      );
     },
     [
       isChatting,
@@ -404,7 +405,7 @@ const ChatBox = (
                 py={4}
                 _hover={{
                   '& .control': {
-                    display: 'flex'
+                    display: item.status === 'finish' ? 'flex' : 'none'
                   }
                 }}
               >
diff --git a/client/src/constants/app.ts b/client/src/constants/app.ts
index ff09156b3..2b4ebd23f 100644
--- a/client/src/constants/app.ts
+++ b/client/src/constants/app.ts
@@ -965,8 +965,8 @@ export const appTemplates: (AppItemType & { avatar: string; intro: string })[] =
         name: '意图识别',
         intro: '可以判断用户问题属于哪方面问题，从而执行不同的操作。',
         type: 'http',
-        url: '/openapi/modules/agent/classifyQuestion',
-        flowType: 'classifyQuestionNode',
+        url: '/openapi/modules/agent/recognizeIntention',
+        flowType: 'recognizeIntention',
         inputs: [
           {
             key: 'systemPrompt',
diff --git a/client/src/constants/data.ts b/client/src/constants/data.ts
deleted file mode 100644
index e72f987fb..000000000
--- a/client/src/constants/data.ts
+++ /dev/null
@@ -1,12 +0,0 @@
-export enum ChatModelEnum {
-  'GPT35' = 'gpt-3.5-turbo',
-  'GPT3516k' = 'gpt-3.5-turbo-16k',
-  'GPT4' = 'gpt-4',
-  'GPT432k' = 'gpt-4-32k'
-}
-
-export const chatModelList = [
-  { label: 'Gpt35-16k', value: ChatModelEnum.GPT3516k },
-  { label: 'Gpt35-4k', value: ChatModelEnum.GPT35 },
-  { label: 'Gpt4-8k', value: ChatModelEnum.GPT4 }
-];
diff --git a/client/src/constants/flow/ModuleTemplate.ts b/client/src/constants/flow/ModuleTemplate.ts
index d945d1e28..5bf0bad98 100644
--- a/client/src/constants/flow/ModuleTemplate.ts
+++ b/client/src/constants/flow/ModuleTemplate.ts
@@ -1,7 +1,7 @@
 import { AppModuleItemTypeEnum, SystemInputEnum, SpecificInputEnum } from '../app';
 import { FlowModuleTypeEnum, FlowInputItemTypeEnum, FlowOutputItemTypeEnum } from './index';
 import type { AppModuleTemplateItemType } from '@/types/app';
-import { chatModelList } from '../data';
+import { chatModelList } from '@/store/static';
 import {
   Input_Template_History,
   Input_Template_TFSwitch,
@@ -96,8 +96,8 @@ export const ChatModule: AppModuleTemplateItemType = {
       key: 'model',
       type: FlowInputItemTypeEnum.select,
       label: '对话模型',
-      value: chatModelList[0].value,
-      list: chatModelList
+      value: chatModelList[0]?.model,
+      list: chatModelList.map((item) => ({ label: item.name, value: item.model }))
     },
     {
       key: 'temperature',
@@ -278,13 +278,13 @@ export const TFSwitchModule: AppModuleTemplateItemType = {
     }
   ]
 };
-export const ClassifyQuestionModule: AppModuleTemplateItemType = {
+export const RecognizeIntentionModule: AppModuleTemplateItemType = {
   logo: '/imgs/module/cq.png',
   name: '意图识别',
   intro: '可以判断用户问题属于哪方面问题，从而执行不同的操作。',
   type: AppModuleItemTypeEnum.http,
-  url: '/openapi/modules/agent/classifyQuestion',
-  flowType: FlowModuleTypeEnum.classifyQuestionNode,
+  url: '/openapi/modules/agent/recognizeIntention',
+  flowType: FlowModuleTypeEnum.recognizeIntention,
   inputs: [
     {
       key: 'systemPrompt',
@@ -348,6 +348,6 @@ export const ModuleTemplates = [
   },
   {
     label: 'Agent',
-    list: [ClassifyQuestionModule]
+    list: [RecognizeIntentionModule]
   }
 ];
diff --git a/client/src/constants/flow/index.ts b/client/src/constants/flow/index.ts
index 4af9271b9..7e79c460c 100644
--- a/client/src/constants/flow/index.ts
+++ b/client/src/constants/flow/index.ts
@@ -26,7 +26,7 @@ export enum FlowModuleTypeEnum {
   kbSearchNode = 'kbSearchNode',
   tfSwitchNode = 'tfSwitchNode',
   answerNode = 'answerNode',
-  classifyQuestionNode = 'classifyQuestionNode'
+  recognizeIntention = 'recognizeIntention'
 }
 
 export const edgeOptions = {
diff --git a/client/src/constants/model.ts b/client/src/constants/model.ts
index f1ddb5551..b8ddf6f47 100644
--- a/client/src/constants/model.ts
+++ b/client/src/constants/model.ts
@@ -1,11 +1,6 @@
-import { getSystemModelList } from '@/api/system';
 import type { ShareChatEditType } from '@/types/app';
 import type { AppSchema } from '@/types/mongoSchema';
 
-export const embeddingModel = 'text-embedding-ada-002';
-export const embeddingPrice = 0.1;
-export type EmbeddingModelType = 'text-embedding-ada-002';
-
 export enum OpenAiChatEnum {
   'GPT35' = 'gpt-3.5-turbo',
   'GPT3516k' = 'gpt-3.5-turbo-16k',
@@ -13,58 +8,6 @@ export enum OpenAiChatEnum {
   'GPT432k' = 'gpt-4-32k'
 }
 
-export type ChatModelType = `${OpenAiChatEnum}`;
-
-export type ChatModelItemType = {
-  chatModel: ChatModelType;
-  name: string;
-  contextMaxToken: number;
-  systemMaxToken: number;
-  maxTemperature: number;
-  price: number;
-};
-
-export const ChatModelMap = {
-  [OpenAiChatEnum.GPT35]: {
-    chatModel: OpenAiChatEnum.GPT35,
-    name: 'Gpt35-4k',
-    contextMaxToken: 4000,
-    systemMaxToken: 2400,
-    maxTemperature: 1.2,
-    price: 1.5
-  },
-  [OpenAiChatEnum.GPT3516k]: {
-    chatModel: OpenAiChatEnum.GPT3516k,
-    name: 'Gpt35-16k',
-    contextMaxToken: 16000,
-    systemMaxToken: 8000,
-    maxTemperature: 1.2,
-    price: 3
-  },
-  [OpenAiChatEnum.GPT4]: {
-    chatModel: OpenAiChatEnum.GPT4,
-    name: 'Gpt4',
-    contextMaxToken: 8000,
-    systemMaxToken: 4000,
-    maxTemperature: 1.2,
-    price: 45
-  },
-  [OpenAiChatEnum.GPT432k]: {
-    chatModel: OpenAiChatEnum.GPT432k,
-    name: 'Gpt4-32k',
-    contextMaxToken: 32000,
-    systemMaxToken: 8000,
-    maxTemperature: 1.2,
-    price: 90
-  }
-};
-
-export const chatModelList: ChatModelItemType[] = [
-  ChatModelMap[OpenAiChatEnum.GPT3516k],
-  ChatModelMap[OpenAiChatEnum.GPT35],
-  ChatModelMap[OpenAiChatEnum.GPT4]
-];
-
 export const defaultApp: AppSchema = {
   _id: '',
   userId: 'userId',
@@ -72,17 +15,6 @@ export const defaultApp: AppSchema = {
   avatar: '/icon/logo.png',
   intro: '',
   updateTime: Date.now(),
-  chat: {
-    relatedKbs: [],
-    searchSimilarity: 0.2,
-    searchLimit: 5,
-    searchEmptyText: '',
-    systemPrompt: '',
-    limitPrompt: '',
-    temperature: 0,
-    maxToken: 4000,
-    chatModel: OpenAiChatEnum.GPT35
-  },
   share: {
     isShare: false,
     isShareDetail: false,
diff --git a/client/src/constants/user.ts b/client/src/constants/user.ts
index 4386331e2..af407832f 100644
--- a/client/src/constants/user.ts
+++ b/client/src/constants/user.ts
@@ -1,9 +1,6 @@
-export enum BillTypeEnum {
-  chat = 'chat',
-  openapiChat = 'openapiChat',
-  QA = 'QA',
-  vector = 'vector',
-  return = 'return'
+export enum BillSourceEnum {
+  fastgpt = 'fastgpt',
+  api = 'api'
 }
 export enum PageTypeEnum {
   login = 'login',
@@ -11,12 +8,9 @@ export enum PageTypeEnum {
   forgetPassword = 'forgetPassword'
 }
 
-export const BillTypeMap: Record<`${BillTypeEnum}`, string> = {
-  [BillTypeEnum.chat]: '对话',
-  [BillTypeEnum.openapiChat]: 'api 对话',
-  [BillTypeEnum.QA]: 'QA拆分',
-  [BillTypeEnum.vector]: '索引生成',
-  [BillTypeEnum.return]: '退款'
+export const BillSourceMap: Record<`${BillSourceEnum}`, string> = {
+  [BillSourceEnum.fastgpt]: 'FastGpt 平台',
+  [BillSourceEnum.api]: 'Api'
 };
 
 export enum PromotionEnum {
diff --git a/client/src/hooks/usePagination.tsx b/client/src/hooks/usePagination.tsx
index 49a8cb64b..516f085ed 100644
--- a/client/src/hooks/usePagination.tsx
+++ b/client/src/hooks/usePagination.tsx
@@ -1,4 +1,4 @@
-import { useRef, useState, useCallback, useLayoutEffect, useMemo, useEffect } from 'react';
+import { useRef, useState, useCallback, useMemo, useEffect } from 'react';
 import type { PagingData } from '../types/index';
 import { IconButton, Flex, Box, Input } from '@chakra-ui/react';
 import { ArrowBackIcon, ArrowForwardIcon } from '@chakra-ui/icons';
@@ -144,7 +144,7 @@ export const usePagination = <T = any,>({
     [data.length, isLoading, mutate, pageNum, total]
   );
 
-  useLayoutEffect(() => {
+  useEffect(() => {
     if (!elementRef.current || type !== 'scroll') return;
 
     const scrolling = throttle((e: Event) => {
diff --git a/client/src/pages/_app.tsx b/client/src/pages/_app.tsx
index 5f58580e9..4f6617cb6 100644
--- a/client/src/pages/_app.tsx
+++ b/client/src/pages/_app.tsx
@@ -1,4 +1,4 @@
-import { useEffect } from 'react';
+import { useEffect, useState } from 'react';
 import type { AppProps } from 'next/app';
 import Script from 'next/script';
 import Head from 'next/head';
@@ -8,9 +8,9 @@ import { theme } from '@/constants/theme';
 import { QueryClient, QueryClientProvider } from '@tanstack/react-query';
 import NProgress from 'nprogress'; //nprogress module
 import Router from 'next/router';
-import { useGlobalStore } from '@/store/global';
 import 'nprogress/nprogress.css';
 import '@/styles/reset.scss';
+import { clientInitData } from '@/store/static';
 
 //Binding events.
 Router.events.on('routeChangeStart', () => NProgress.start());
@@ -29,13 +29,15 @@ const queryClient = new QueryClient({
 });
 
 function App({ Component, pageProps }: AppProps) {
-  const {
-    loadInitData,
-    initData: { googleVerKey, baiduTongji }
-  } = useGlobalStore();
+  const [googleVerKey, setGoogleVerKey] = useState<string>();
+  const [baiduTongji, setBaiduTongji] = useState<string>();
 
   useEffect(() => {
-    loadInitData();
+    (async () => {
+      const { googleVerKey, baiduTongji } = await clientInitData();
+      setGoogleVerKey(googleVerKey);
+      setBaiduTongji(baiduTongji);
+    })();
   }, []);
 
   return (
@@ -53,7 +55,7 @@ function App({ Component, pageProps }: AppProps) {
       <Script src="/js/qrcode.min.js" strategy="lazyOnload"></Script>
       <Script src="/js/pdf.js" strategy="lazyOnload"></Script>
       <Script src="/js/html2pdf.bundle.min.js" strategy="lazyOnload"></Script>
-      {baiduTongji && <Script src="/js/baidutongji.js" strategy="lazyOnload"></Script>}
+      {baiduTongji && <Script src={baiduTongji} strategy="lazyOnload"></Script>}
       {googleVerKey && (
         <>
           <Script
@@ -75,5 +77,4 @@ function App({ Component, pageProps }: AppProps) {
   );
 }
 
-// @ts-ignore
 export default App;
diff --git a/client/src/pages/api/chat/chatTest.ts b/client/src/pages/api/chat/chatTest.ts
index 18bfbfb24..5ebe3cb4f 100644
--- a/client/src/pages/api/chat/chatTest.ts
+++ b/client/src/pages/api/chat/chatTest.ts
@@ -8,6 +8,8 @@ import { type ChatCompletionRequestMessage } from 'openai';
 import { AppModuleItemType } from '@/types/app';
 import { dispatchModules } from '../openapi/v1/chat/completions';
 import { gptMessage2ChatType } from '@/utils/adapt';
+import { createTaskBill, delTaskBill, finishTaskBill } from '@/service/events/pushBill';
+import { BillSourceEnum } from '@/constants/user';
 
 export type MessageItemType = ChatCompletionRequestMessage & { _id?: string };
 export type Props = {
@@ -15,10 +17,8 @@ export type Props = {
   prompt: string;
   modules: AppModuleItemType[];
   variables: Record<string, any>;
-};
-export type ChatResponseType = {
-  newChatId: string;
-  quoteLen?: number;
+  appId: string;
+  appName: string;
 };
 
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
@@ -30,8 +30,8 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
     res.end();
   });
 
-  let { modules = [], history = [], prompt, variables = {} } = req.body as Props;
-
+  let { modules = [], history = [], prompt, variables = {}, appName, appId } = req.body as Props;
+  let billId = '';
   try {
     if (!history || !modules || !prompt) {
       throw new Error('Prams Error');
@@ -45,6 +45,13 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
     /* user auth */
     const { userId } = await authUser({ req });
 
+    billId = await createTaskBill({
+      userId,
+      appName,
+      appId,
+      source: BillSourceEnum.fastgpt
+    });
+
     /* start process */
     const { responseData } = await dispatchModules({
       res,
@@ -54,7 +61,8 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
         history: gptMessage2ChatType(history),
         userChatInput: prompt
       },
-      stream: true
+      stream: true,
+      billId
     });
 
     sseResponse({
@@ -70,7 +78,11 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
     res.end();
 
     // bill
+    finishTaskBill({
+      billId
+    });
   } catch (err: any) {
+    delTaskBill(billId);
     res.status(500);
     sseErrRes(res, err);
     res.end();
diff --git a/client/src/pages/api/chat/saveChat.ts b/client/src/pages/api/chat/saveChat.ts
index 2a7fea8c9..cc7f9d1f4 100644
--- a/client/src/pages/api/chat/saveChat.ts
+++ b/client/src/pages/api/chat/saveChat.ts
@@ -53,14 +53,6 @@ export async function saveChat({
   await connectToDatabase();
   const { app } = await authApp({ appId, userId, authOwner: false });
 
-  const content = prompts.map((item) => ({
-    _id: item._id,
-    obj: item.obj,
-    value: item.value,
-    systemPrompt: item.systemPrompt || '',
-    quote: item.quote || []
-  }));
-
   if (String(app.userId) === userId) {
     await App.findByIdAndUpdate(appId, {
       updateTime: new Date()
@@ -73,12 +65,11 @@ export async function saveChat({
           Chat.findByIdAndUpdate(historyId, {
             $push: {
               content: {
-                $each: content
+                $each: prompts
               }
             },
             variables,
-            title: content[0].value.slice(0, 20),
-            latestChat: content[1].value,
+            title: prompts[0].value.slice(0, 20),
             updateTime: new Date()
           }).then(() => ({
             newHistoryId: ''
@@ -90,9 +81,8 @@ export async function saveChat({
             userId,
             appId,
             variables,
-            content,
-            title: content[0].value.slice(0, 20),
-            latestChat: content[1].value
+            content: prompts,
+            title: prompts[0].value.slice(0, 20)
           }).then((res) => ({
             newHistoryId: String(res._id)
           }))
diff --git a/client/src/pages/api/openapi/kb/appKbSearch.ts b/client/src/pages/api/openapi/kb/appKbSearch.ts
deleted file mode 100644
index 43f7cb2d3..000000000
--- a/client/src/pages/api/openapi/kb/appKbSearch.ts
+++ /dev/null
@@ -1,186 +0,0 @@
-import type { NextApiRequest, NextApiResponse } from 'next';
-import { jsonRes } from '@/service/response';
-import { authUser } from '@/service/utils/auth';
-import { PgClient } from '@/service/pg';
-import { withNextCors } from '@/service/utils/tools';
-import type { ChatItemType } from '@/types/chat';
-import type { AppSchema } from '@/types/mongoSchema';
-import { authApp } from '@/service/utils/auth';
-import { ChatModelMap } from '@/constants/model';
-import { ChatRoleEnum } from '@/constants/chat';
-import { openaiEmbedding } from '../plugin/openaiEmbedding';
-import { modelToolMap } from '@/utils/plugin';
-
-export type QuoteItemType = {
-  id: string;
-  q: string;
-  a: string;
-  source?: string;
-};
-type Props = {
-  prompts: ChatItemType[];
-  similarity: number;
-  limit: number;
-  appId: string;
-};
-type Response = {
-  rawSearch: QuoteItemType[];
-  userSystemPrompt: {
-    obj: ChatRoleEnum;
-    value: string;
-  }[];
-  userLimitPrompt: {
-    obj: ChatRoleEnum;
-    value: string;
-  }[];
-  quotePrompt: {
-    obj: ChatRoleEnum;
-    value: string;
-  };
-};
-
-export default withNextCors(async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
-  try {
-    const { userId } = await authUser({ req });
-
-    if (!userId) {
-      throw new Error('userId is empty');
-    }
-
-    const { prompts, similarity, limit, appId } = req.body as Props;
-
-    if (!similarity || !Array.isArray(prompts) || !appId) {
-      throw new Error('params is error');
-    }
-
-    // auth app
-    const { app } = await authApp({
-      appId,
-      userId
-    });
-
-    const result = await appKbSearch({
-      app,
-      userId,
-      fixedQuote: [],
-      prompt: prompts[prompts.length - 1],
-      similarity,
-      limit
-    });
-
-    jsonRes<Response>(res, {
-      data: result
-    });
-  } catch (err) {
-    console.log(err);
-    jsonRes(res, {
-      code: 500,
-      error: err
-    });
-  }
-});
-
-export async function appKbSearch({
-  app,
-  userId,
-  fixedQuote = [],
-  prompt,
-  similarity = 0.8,
-  limit = 5
-}: {
-  app: AppSchema;
-  userId: string;
-  fixedQuote?: QuoteItemType[];
-  prompt: ChatItemType;
-  similarity: number;
-  limit: number;
-}): Promise<Response> {
-  const modelConstantsData = ChatModelMap[app.chat.chatModel];
-
-  // get vector
-  const promptVector = await openaiEmbedding({
-    userId,
-    input: [prompt.value]
-  });
-
-  // search kb
-  const res: any = await PgClient.query(
-    `BEGIN;
-    SET LOCAL ivfflat.probes = ${global.systemEnv.pgIvfflatProbe || 10};
-    select id,q,a,source from modelData where kb_id IN (${app.chat.relatedKbs
-      .map((item) => `'${item}'`)
-      .join(',')}) AND vector <#> '[${promptVector[0]}]' < -${similarity} order by vector <#> '[${
-      promptVector[0]
-    }]' limit ${limit};
-    COMMIT;`
-  );
-
-  const searchRes: QuoteItemType[] = res?.[2]?.rows || [];
-
-  // filter same search result
-  const idSet = new Set<string>();
-  const filterSearch = [
-    ...searchRes.slice(0, 3),
-    ...fixedQuote.slice(0, 2),
-    ...searchRes.slice(3),
-    ...fixedQuote.slice(2, Math.floor(fixedQuote.length * 0.4))
-  ].filter((item) => {
-    if (idSet.has(item.id)) {
-      return false;
-    }
-    idSet.add(item.id);
-    return true;
-  });
-
-  // 计算固定提示词的 token 数量
-  const userSystemPrompt = app.chat.systemPrompt // user system prompt
-    ? [
-        {
-          obj: ChatRoleEnum.System,
-          value: app.chat.systemPrompt
-        }
-      ]
-    : [];
-  const userLimitPrompt = [
-    {
-      obj: ChatRoleEnum.Human,
-      value: app.chat.limitPrompt
-        ? app.chat.limitPrompt
-        : `知识库是关于 ${app.name} 的内容，参考知识库回答问题。与 "${app.name}" 无关内容，直接回复: "我不知道"。`
-    }
-  ];
-
-  const fixedSystemTokens = modelToolMap.countTokens({
-    model: app.chat.chatModel,
-    messages: [...userSystemPrompt, ...userLimitPrompt]
-  });
-
-  // filter part quote by maxToken
-  const sliceResult = modelToolMap
-    .tokenSlice({
-      model: app.chat.chatModel,
-      maxToken: modelConstantsData.systemMaxToken - fixedSystemTokens,
-      messages: filterSearch.map((item, i) => ({
-        obj: ChatRoleEnum.System,
-        value: `${i + 1}: [${item.q}\n${item.a}]`
-      }))
-    })
-    .map((item) => item.value)
-    .join('\n')
-    .trim();
-
-  // slice filterSearch
-  const rawSearch = filterSearch.slice(0, sliceResult.length);
-
-  const quoteText = sliceResult ? `知识库:\n${sliceResult}` : '';
-
-  return {
-    rawSearch,
-    userSystemPrompt,
-    userLimitPrompt,
-    quotePrompt: {
-      obj: ChatRoleEnum.System,
-      value: quoteText
-    }
-  };
-}
diff --git a/client/src/pages/api/openapi/kb/pushData.ts b/client/src/pages/api/openapi/kb/pushData.ts
index d3fad4f7a..1b58a3097 100644
--- a/client/src/pages/api/openapi/kb/pushData.ts
+++ b/client/src/pages/api/openapi/kb/pushData.ts
@@ -15,6 +15,7 @@ type DateItemType = { a: string; q: string; source?: string };
 export type Props = {
   kbId: string;
   data: DateItemType[];
+  model: string;
   mode: `${TrainingModeEnum}`;
   prompt?: string;
 };
@@ -25,14 +26,14 @@ export type Response = {
 
 const modeMaxToken = {
   [TrainingModeEnum.index]: 6000,
-  [TrainingModeEnum.qa]: 10000
+  [TrainingModeEnum.qa]: 12000
 };
 
 export default withNextCors(async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
   try {
-    const { kbId, data, mode, prompt } = req.body as Props;
+    const { kbId, data, mode, prompt, model } = req.body as Props;
 
-    if (!kbId || !Array.isArray(data)) {
+    if (!kbId || !Array.isArray(data) || !model) {
       throw new Error('缺少参数');
     }
     await connectToDatabase();
@@ -46,7 +47,8 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
         data,
         userId,
         mode,
-        prompt
+        prompt,
+        model
       })
     });
   } catch (err) {
@@ -62,7 +64,8 @@ export async function pushDataToKb({
   kbId,
   data,
   mode,
-  prompt
+  prompt,
+  model
 }: { userId: string } & Props): Promise<Response> {
   await authKb({
     userId,
@@ -79,7 +82,7 @@ export async function pushDataToKb({
     if (mode === TrainingModeEnum.qa) {
       // count token
       const token = modelToolMap.countTokens({
-        model: OpenAiChatEnum.GPT3516k,
+        model: 'gpt-3.5-turbo-16k',
         messages: [{ obj: 'System', value: item.q }]
       });
       if (token > modeMaxToken[TrainingModeEnum.qa]) {
@@ -144,6 +147,7 @@ export async function pushDataToKb({
     insertData.map((item) => ({
       q: item.q,
       a: item.a,
+      model,
       source: item.source,
       userId,
       kbId,
diff --git a/client/src/pages/api/openapi/kb/searchTest.ts b/client/src/pages/api/openapi/kb/searchTest.ts
index b1aa608f8..d2374ac16 100644
--- a/client/src/pages/api/openapi/kb/searchTest.ts
+++ b/client/src/pages/api/openapi/kb/searchTest.ts
@@ -3,7 +3,7 @@ import { jsonRes } from '@/service/response';
 import { authUser } from '@/service/utils/auth';
 import { PgClient } from '@/service/pg';
 import { withNextCors } from '@/service/utils/tools';
-import { openaiEmbedding } from '../plugin/openaiEmbedding';
+import { getVector } from '../plugin/vector';
 import type { KbTestItemType } from '@/types/plugin';
 
 export type Props = {
@@ -27,7 +27,7 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
       throw new Error('缺少用户ID');
     }
 
-    const vector = await openaiEmbedding({
+    const vector = await getVector({
       userId,
       input: [text]
     });
diff --git a/client/src/pages/api/openapi/kb/updateData.ts b/client/src/pages/api/openapi/kb/updateData.ts
index f38fb6240..aebe83670 100644
--- a/client/src/pages/api/openapi/kb/updateData.ts
+++ b/client/src/pages/api/openapi/kb/updateData.ts
@@ -3,7 +3,7 @@ import { jsonRes } from '@/service/response';
 import { authUser } from '@/service/utils/auth';
 import { PgClient } from '@/service/pg';
 import { withNextCors } from '@/service/utils/tools';
-import { openaiEmbedding } from '../plugin/openaiEmbedding';
+import { getVector } from '../plugin/vector';
 
 export default withNextCors(async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
   try {
@@ -19,7 +19,7 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
     // get vector
     const vector = await (async () => {
       if (q) {
-        return openaiEmbedding({
+        return getVector({
           userId,
           input: [q]
         });
diff --git a/client/src/pages/api/openapi/modules/agent/extract.ts b/client/src/pages/api/openapi/modules/agent/extract.ts
index a570d7911..84b99ec69 100644
--- a/client/src/pages/api/openapi/modules/agent/extract.ts
+++ b/client/src/pages/api/openapi/modules/agent/extract.ts
@@ -6,12 +6,12 @@ import { ChatContextFilter } from '@/service/utils/chat/index';
 import type { ChatItemType } from '@/types/chat';
 import { ChatRoleEnum } from '@/constants/chat';
 import { getOpenAIApi, axiosConfig } from '@/service/ai/openai';
-import type { ClassifyQuestionAgentItemType } from '@/types/app';
+import type { RecognizeIntentionAgentItemType } from '@/types/app';
 
 export type Props = {
   history?: ChatItemType[];
   userChatInput: string;
-  agents: ClassifyQuestionAgentItemType[];
+  agents: RecognizeIntentionAgentItemType[];
   description: string;
 };
 export type Response = { history: ChatItemType[] };
diff --git a/client/src/pages/api/openapi/modules/agent/classifyQuestion.ts b/client/src/pages/api/openapi/modules/agent/recognizeIntention.ts
similarity index 77%
rename from client/src/pages/api/openapi/modules/agent/classifyQuestion.ts
rename to client/src/pages/api/openapi/modules/agent/recognizeIntention.ts
index 9c5af1e8b..dfb10e1b9 100644
--- a/client/src/pages/api/openapi/modules/agent/classifyQuestion.ts
+++ b/client/src/pages/api/openapi/modules/agent/recognizeIntention.ts
@@ -6,29 +6,30 @@ import { ChatContextFilter } from '@/service/utils/chat/index';
 import type { ChatItemType } from '@/types/chat';
 import { ChatRoleEnum } from '@/constants/chat';
 import { getOpenAIApi, axiosConfig } from '@/service/ai/openai';
-import type { ClassifyQuestionAgentItemType } from '@/types/app';
+import type { RecognizeIntentionAgentItemType } from '@/types/app';
+import { countModelPrice, pushTaskBillListItem } from '@/service/events/pushBill';
 
 export type Props = {
   systemPrompt?: string;
   history?: ChatItemType[];
   userChatInput: string;
-  agents: ClassifyQuestionAgentItemType[];
+  agents: RecognizeIntentionAgentItemType[];
+  billId?: string;
 };
 export type Response = { history: ChatItemType[] };
 
-const agentModel = 'gpt-3.5-turbo-16k';
+const agentModel = 'gpt-3.5-turbo';
 const agentFunName = 'agent_user_question';
 
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
   try {
-    let { systemPrompt, agents, history = [], userChatInput } = req.body as Props;
+    let { userChatInput } = req.body as Props;
 
-    const response = await classifyQuestion({
-      systemPrompt,
-      history,
-      userChatInput,
-      agents
-    });
+    if (!userChatInput) {
+      throw new Error('userChatInput is empty');
+    }
+
+    const response = await classifyQuestion(req.body);
 
     jsonRes(res, {
       data: response
@@ -46,7 +47,8 @@ export async function classifyQuestion({
   agents,
   systemPrompt,
   history = [],
-  userChatInput
+  userChatInput,
+  billId
 }: Props) {
   const messages: ChatItemType[] = [
     ...(systemPrompt
@@ -106,8 +108,19 @@ export async function classifyQuestion({
   if (!arg.type) {
     throw new Error('');
   }
+
+  const totalTokens = response.data.usage?.total_tokens || 0;
+
+  await pushTaskBillListItem({
+    billId,
+    moduleName: 'Recognize Intention',
+    amount: countModelPrice({ model: agentModel, tokens: totalTokens }),
+    model: agentModel,
+    tokenLen: totalTokens
+  });
+
   console.log(
-    '意图结果',
+    'CQ',
     agents.findIndex((item) => item.key === arg.type)
   );
 
diff --git a/client/src/pages/api/openapi/modules/chat/gpt.ts b/client/src/pages/api/openapi/modules/chat/gpt.ts
index f8ac885c7..b12e8ae4b 100644
--- a/client/src/pages/api/openapi/modules/chat/gpt.ts
+++ b/client/src/pages/api/openapi/modules/chat/gpt.ts
@@ -1,9 +1,9 @@
 // Next.js API route support: https://nextjs.org/docs/api-routes/introduction
 import type { NextApiRequest, NextApiResponse } from 'next';
-import { jsonRes } from '@/service/response';
+import { jsonRes, sseErrRes } from '@/service/response';
 import { sseResponse } from '@/service/utils/tools';
-import { ChatModelMap, OpenAiChatEnum } from '@/constants/model';
-import { adaptChatItem_openAI } from '@/utils/plugin/openai';
+import { OpenAiChatEnum } from '@/constants/model';
+import { adaptChatItem_openAI, countOpenAIToken } from '@/utils/plugin/openai';
 import { modelToolMap } from '@/utils/plugin';
 import { ChatContextFilter } from '@/service/utils/chat/index';
 import type { ChatItemType } from '@/types/chat';
@@ -11,6 +11,8 @@ import { ChatRoleEnum, sseResponseEventEnum } from '@/constants/chat';
 import { parseStreamChunk, textAdaptGptResponse } from '@/utils/adapt';
 import { getOpenAIApi, axiosConfig } from '@/service/ai/openai';
 import { SpecificInputEnum } from '@/constants/app';
+import { getChatModel } from '@/service/utils/data';
+import { countModelPrice, pushTaskBillListItem } from '@/service/events/pushBill';
 
 export type Props = {
   model: `${OpenAiChatEnum}`;
@@ -22,39 +24,28 @@ export type Props = {
   quotePrompt?: string;
   systemPrompt?: string;
   limitPrompt?: string;
+  billId?: string;
 };
-export type Response = { [SpecificInputEnum.answerText]: string };
+export type Response = { [SpecificInputEnum.answerText]: string; totalTokens: number };
 
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
+  let { model, temperature = 0, stream } = req.body as Props;
   try {
-    let {
-      model,
-      stream = false,
-      temperature = 0,
-      maxToken = 4000,
-      history = [],
-      quotePrompt,
-      userChatInput,
-      systemPrompt,
-      limitPrompt
-    } = req.body as Props;
-
     // temperature adapt
-    const modelConstantsData = ChatModelMap[model];
+    const modelConstantsData = getChatModel(model);
+
+    if (!modelConstantsData) {
+      throw new Error('The chat model is undefined');
+    }
+
     // FastGpt temperature range: 1~10
     temperature = +(modelConstantsData.maxTemperature * (temperature / 10)).toFixed(2);
 
     const response = await chatCompletion({
+      ...req.body,
       res,
       model,
-      temperature,
-      maxToken,
-      stream,
-      history,
-      userChatInput,
-      systemPrompt,
-      limitPrompt,
-      quotePrompt
+      temperature
     });
 
     if (stream) {
@@ -70,25 +61,32 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
       });
     }
   } catch (err) {
-    jsonRes(res, {
-      code: 500,
-      error: err
-    });
+    if (stream) {
+      res.status(500);
+      sseErrRes(res, err);
+      res.end();
+    } else {
+      jsonRes(res, {
+        code: 500,
+        error: err
+      });
+    }
   }
 }
 
 /* request openai chat */
 export async function chatCompletion({
   res,
-  model = OpenAiChatEnum.GPT35,
-  temperature,
+  model,
+  temperature = 0,
   maxToken = 4000,
-  stream,
+  stream = false,
   history = [],
-  quotePrompt,
+  quotePrompt = '',
   userChatInput,
-  systemPrompt,
-  limitPrompt
+  systemPrompt = '',
+  limitPrompt = '',
+  billId
 }: Props & { res: NextApiResponse }): Promise<Response> {
   const messages: ChatItemType[] = [
     ...(quotePrompt
@@ -121,7 +119,7 @@ export async function chatCompletion({
       value: userChatInput
     }
   ];
-  const modelTokenLimit = ChatModelMap[model]?.contextMaxToken || 4000;
+  const modelTokenLimit = getChatModel(model)?.contextMaxToken || 4000;
 
   const filterMessages = ChatContextFilter({
     model,
@@ -157,37 +155,47 @@ export async function chatCompletion({
     }
   );
 
-  const { answer } = await (async () => {
+  const { answer, totalTokens } = await (async () => {
     if (stream) {
       // sse response
       const { answer } = await streamResponse({ res, response });
       // count tokens
-      // const finishMessages = filterMessages.concat({
-      //   obj: ChatRoleEnum.AI,
-      //   value: answer
-      // });
+      const finishMessages = filterMessages.concat({
+        obj: ChatRoleEnum.AI,
+        value: answer
+      });
 
-      // const totalTokens = modelToolMap[model].countTokens({
-      //   messages: finishMessages
-      // });
+      const totalTokens = countOpenAIToken({
+        messages: finishMessages,
+        model: 'gpt-3.5-turbo-16k'
+      });
 
       return {
-        answer
-        // totalTokens
+        answer,
+        totalTokens
       };
     } else {
       const answer = stream ? '' : response.data.choices?.[0].message?.content || '';
-      // const totalTokens = stream ? 0 : response.data.usage?.total_tokens || 0;
+      const totalTokens = stream ? 0 : response.data.usage?.total_tokens || 0;
 
       return {
-        answer
-        // totalTokens
+        answer,
+        totalTokens
       };
     }
   })();
 
+  await pushTaskBillListItem({
+    billId,
+    moduleName: 'AI Chat',
+    amount: countModelPrice({ model, tokens: totalTokens }),
+    model,
+    tokenLen: totalTokens
+  });
+
   return {
-    answerText: answer
+    answerText: answer,
+    totalTokens
   };
 }
 
diff --git a/client/src/pages/api/openapi/modules/kb/search.ts b/client/src/pages/api/openapi/modules/kb/search.ts
index f9c934a7e..eee998c77 100644
--- a/client/src/pages/api/openapi/modules/kb/search.ts
+++ b/client/src/pages/api/openapi/modules/kb/search.ts
@@ -4,8 +4,9 @@ import { PgClient } from '@/service/pg';
 import { withNextCors } from '@/service/utils/tools';
 import type { ChatItemType } from '@/types/chat';
 import { ChatRoleEnum } from '@/constants/chat';
-import { openaiEmbedding_system } from '../../plugin/openaiEmbedding';
 import { modelToolMap } from '@/utils/plugin';
+import { getVector } from '../../plugin/vector';
+import { countModelPrice, pushTaskBillListItem } from '@/service/events/pushBill';
 
 export type QuoteItemType = {
   id: string;
@@ -21,6 +22,7 @@ type Props = {
   maxToken: number;
   userChatInput: string;
   stream?: boolean;
+  billId?: string;
 };
 type Response = {
   rawSearch: QuoteItemType[];
@@ -30,25 +32,15 @@ type Response = {
 
 export default withNextCors(async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
   try {
-    const {
-      kb_ids = [],
-      history = [],
-      similarity,
-      limit,
-      maxToken,
-      userChatInput
-    } = req.body as Props;
+    const { kb_ids = [], userChatInput } = req.body as Props;
 
-    if (!similarity || !Array.isArray(kb_ids)) {
+    if (!userChatInput || !Array.isArray(kb_ids)) {
       throw new Error('params is error');
     }
 
     const result = await kbSearch({
+      ...req.body,
       kb_ids,
-      history,
-      similarity,
-      limit,
-      maxToken,
       userChatInput
     });
 
@@ -70,7 +62,8 @@ export async function kbSearch({
   similarity = 0.8,
   limit = 5,
   maxToken = 2500,
-  userChatInput
+  userChatInput,
+  billId
 }: Props): Promise<Response> {
   if (kb_ids.length === 0)
     return {
@@ -78,22 +71,34 @@ export async function kbSearch({
       rawSearch: [],
       quotePrompt: undefined
     };
+
   // get vector
-  const promptVector = await openaiEmbedding_system({
+  const vectorModel = global.vectorModels[0].model;
+  const { vectors, tokenLen } = await getVector({
+    model: vectorModel,
     input: [userChatInput]
   });
 
   // search kb
-  const res: any = await PgClient.query(
-    `BEGIN;
+  const [res]: any = await Promise.all([
+    PgClient.query(
+      `BEGIN;
     SET LOCAL ivfflat.probes = ${global.systemEnv.pgIvfflatProbe || 10};
     select id,q,a,source from modelData where kb_id IN (${kb_ids
       .map((item) => `'${item}'`)
-      .join(',')}) AND vector <#> '[${promptVector[0]}]' < -${similarity} order by vector <#> '[${
-      promptVector[0]
-    }]' limit ${limit};
+      .join(',')}) AND vector <#> '[${vectors[0]}]' < -${similarity} order by vector <#> '[${
+        vectors[0]
+      }]' limit ${limit};
     COMMIT;`
-  );
+    ),
+    pushTaskBillListItem({
+      billId,
+      moduleName: 'Vector Generate',
+      amount: countModelPrice({ model: vectorModel, tokens: tokenLen }),
+      model: vectorModel,
+      tokenLen
+    })
+  ]);
 
   const searchRes: QuoteItemType[] = res?.[2]?.rows || [];
 
diff --git a/client/src/pages/api/openapi/plugin/openaiEmbedding.ts b/client/src/pages/api/openapi/plugin/openaiEmbedding.ts
deleted file mode 100644
index 841100d80..000000000
--- a/client/src/pages/api/openapi/plugin/openaiEmbedding.ts
+++ /dev/null
@@ -1,115 +0,0 @@
-import type { NextApiRequest, NextApiResponse } from 'next';
-import { jsonRes } from '@/service/response';
-import { authUser, getApiKey, getSystemOpenAiKey } from '@/service/utils/auth';
-import { withNextCors } from '@/service/utils/tools';
-import { getOpenAIApi } from '@/service/utils/chat/openai';
-import { embeddingModel } from '@/constants/model';
-import { axiosConfig } from '@/service/utils/tools';
-import { pushGenerateVectorBill } from '@/service/events/pushBill';
-import { OpenAiChatEnum } from '@/constants/model';
-
-type Props = {
-  input: string[];
-};
-type Response = number[][];
-
-export default withNextCors(async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
-  try {
-    const { userId } = await authUser({ req });
-    let { input } = req.query as Props;
-
-    if (!Array.isArray(input)) {
-      throw new Error('缺少参数');
-    }
-
-    jsonRes<Response>(res, {
-      data: await openaiEmbedding({ userId, input, mustPay: true })
-    });
-  } catch (err) {
-    console.log(err);
-    jsonRes(res, {
-      code: 500,
-      error: err
-    });
-  }
-});
-
-export async function openaiEmbedding({
-  userId,
-  input,
-  mustPay = false
-}: { userId: string; mustPay?: boolean } & Props) {
-  const { userOpenAiKey, systemAuthKey } = await getApiKey({
-    model: 'gpt-3.5-turbo',
-    userId,
-    mustPay
-  });
-  const apiKey = userOpenAiKey || systemAuthKey;
-
-  // 获取 chatAPI
-  const chatAPI = getOpenAIApi(apiKey);
-
-  // 把输入的内容转成向量
-  const result = await chatAPI
-    .createEmbedding(
-      {
-        model: embeddingModel,
-        input
-      },
-      {
-        timeout: 60000,
-        ...axiosConfig(apiKey)
-      }
-    )
-    .then((res) => {
-      if (!res.data?.usage?.total_tokens) {
-        // @ts-ignore
-        return Promise.reject(res.data?.error?.message || 'Embedding Error');
-      }
-      return {
-        tokenLen: res.data.usage.total_tokens || 0,
-        vectors: res.data.data.map((item) => item.embedding)
-      };
-    });
-
-  pushGenerateVectorBill({
-    isPay: !userOpenAiKey,
-    userId,
-    text: input.join(''),
-    tokenLen: result.tokenLen
-  });
-
-  return result.vectors;
-}
-
-export async function openaiEmbedding_system({ input }: Props) {
-  const apiKey = getSystemOpenAiKey();
-
-  // 获取 chatAPI
-  const chatAPI = getOpenAIApi(apiKey);
-
-  // 把输入的内容转成向量
-  const result = await chatAPI
-    .createEmbedding(
-      {
-        model: embeddingModel,
-        input
-      },
-      {
-        timeout: 20000,
-        ...axiosConfig(apiKey)
-      }
-    )
-    .then((res) => {
-      if (!res.data?.usage?.total_tokens) {
-        // @ts-ignore
-        return Promise.reject(res.data?.error?.message || 'Embedding Error');
-      }
-      return {
-        tokenLen: res.data.usage.total_tokens || 0,
-        vectors: res.data.data.map((item) => item.embedding)
-      };
-    });
-
-  return result.vectors;
-}
diff --git a/client/src/pages/api/openapi/plugin/vector.ts b/client/src/pages/api/openapi/plugin/vector.ts
new file mode 100644
index 000000000..29b592f08
--- /dev/null
+++ b/client/src/pages/api/openapi/plugin/vector.ts
@@ -0,0 +1,79 @@
+import type { NextApiRequest, NextApiResponse } from 'next';
+import { jsonRes } from '@/service/response';
+import { authBalanceByUid, authUser } from '@/service/utils/auth';
+import { withNextCors } from '@/service/utils/tools';
+import { getOpenAIApi, axiosConfig } from '@/service/ai/openai';
+import { pushGenerateVectorBill } from '@/service/events/pushBill';
+
+type Props = {
+  model: string;
+  input: string[];
+};
+type Response = {
+  tokenLen: number;
+  vectors: number[][];
+};
+
+export default withNextCors(async function handler(req: NextApiRequest, res: NextApiResponse<any>) {
+  try {
+    const { userId } = await authUser({ req });
+    let { input, model } = req.query as Props;
+
+    if (!Array.isArray(input)) {
+      throw new Error('缺少参数');
+    }
+
+    jsonRes<Response>(res, {
+      data: await getVector({ userId, input, model })
+    });
+  } catch (err) {
+    console.log(err);
+    jsonRes(res, {
+      code: 500,
+      error: err
+    });
+  }
+});
+
+export async function getVector({
+  model = 'text-embedding-ada-002',
+  userId,
+  input
+}: { userId?: string } & Props) {
+  userId && (await authBalanceByUid(userId));
+
+  // 获取 chatAPI
+  const chatAPI = getOpenAIApi();
+
+  // 把输入的内容转成向量
+  const result = await chatAPI
+    .createEmbedding(
+      {
+        model,
+        input
+      },
+      {
+        timeout: 60000,
+        ...axiosConfig()
+      }
+    )
+    .then((res) => {
+      if (!res.data?.usage?.total_tokens) {
+        // @ts-ignore
+        return Promise.reject(res.data?.error?.message || 'Embedding Error');
+      }
+      return {
+        tokenLen: res.data.usage.total_tokens || 0,
+        vectors: res.data.data.map((item) => item.embedding)
+      };
+    });
+
+  userId &&
+    pushGenerateVectorBill({
+      userId,
+      tokenLen: result.tokenLen,
+      model
+    });
+
+  return result;
+}
diff --git a/client/src/pages/api/openapi/v1/chat/completions.ts b/client/src/pages/api/openapi/v1/chat/completions.ts
index adee2724f..838eb86bd 100644
--- a/client/src/pages/api/openapi/v1/chat/completions.ts
+++ b/client/src/pages/api/openapi/v1/chat/completions.ts
@@ -15,8 +15,8 @@ import { Types } from 'mongoose';
 import { moduleFetch } from '@/service/api/request';
 import { AppModuleItemType, RunningModuleItemType } from '@/types/app';
 import { FlowInputItemTypeEnum } from '@/constants/flow';
-import { pushChatBill } from '@/service/events/pushBill';
-import { BillTypeEnum } from '@/constants/user';
+import { finishTaskBill, createTaskBill } from '@/service/events/pushBill';
+import { BillSourceEnum } from '@/constants/user';
 
 export type MessageItemType = ChatCompletionRequestMessage & { _id?: string };
 type FastGptWebChatProps = {
@@ -108,6 +108,13 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
       res.setHeader('newHistoryId', String(newHistoryId));
     }
 
+    const billId = await createTaskBill({
+      userId,
+      appName: app.name,
+      appId,
+      source: BillSourceEnum.fastgpt
+    });
+
     /* start process */
     const { responseData, answerText } = await dispatchModules({
       res,
@@ -117,7 +124,8 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
         history: prompts,
         userChatInput: prompt.value
       },
-      stream
+      stream,
+      billId: ''
     });
 
     // save chat
@@ -171,14 +179,9 @@ export default withNextCors(async function handler(req: NextApiRequest, res: Nex
       });
     }
 
-    pushChatBill({
-      isPay: true,
-      chatModel: 'gpt-3.5-turbo',
-      userId,
-      appId,
-      textLen: 1,
-      tokens: 100,
-      type: BillTypeEnum.chat
+    // bill
+    finishTaskBill({
+      billId
     });
   } catch (err: any) {
     if (stream) {
@@ -199,18 +202,21 @@ export async function dispatchModules({
   modules,
   params = {},
   variables = {},
-  stream = false
+  stream = false,
+  billId
 }: {
   res: NextApiResponse;
   modules: AppModuleItemType[];
   params?: Record<string, any>;
   variables?: Record<string, any>;
+  billId: string;
   stream?: boolean;
 }) {
   const runningModules = loadModules(modules, variables);
-  let storeData: Record<string, any> = {};
-  let responseData: Record<string, any> = {};
-  let answerText = '';
+
+  let storeData: Record<string, any> = {}; // after module used
+  let responseData: Record<string, any> = {}; // response request and save to database
+  let answerText = ''; // AI answer
 
   function pushStore({
     isResponse = false,
@@ -327,6 +333,7 @@ export async function dispatchModules({
       });
       const data = {
         stream,
+        billId,
         ...params
       };
 
diff --git a/client/src/pages/api/system/getInitData.ts b/client/src/pages/api/system/getInitData.ts
index 46ce3daee..1fb233ef9 100644
--- a/client/src/pages/api/system/getInitData.ts
+++ b/client/src/pages/api/system/getInitData.ts
@@ -1,19 +1,114 @@
 // Next.js API route support: https://nextjs.org/docs/api-routes/introduction
 import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
+import {
+  type QAModelItemType,
+  type ChatModelItemType,
+  type VectorModelItemType
+} from '@/types/model';
+import { readFileSync } from 'fs';
 
 export type InitDateResponse = {
   beianText: string;
   googleVerKey: string;
-  baiduTongji: boolean;
+  baiduTongji: string;
+  chatModels: ChatModelItemType[];
+  qaModels: QAModelItemType[];
+  vectorModels: VectorModelItemType[];
+};
+
+const defaultmodels = {
+  'Gpt35-4k': {
+    model: 'gpt-3.5-turbo',
+    name: 'Gpt35-4k',
+    contextMaxToken: 4000,
+    systemMaxToken: 2400,
+    maxTemperature: 1.2,
+    price: 1.5
+  },
+  'Gpt35-16k': {
+    model: 'gpt-3.5-turbo',
+    name: 'Gpt35-16k',
+    contextMaxToken: 16000,
+    systemMaxToken: 8000,
+    maxTemperature: 1.2,
+    price: 3
+  },
+  Gpt4: {
+    model: 'gpt-4',
+    name: 'Gpt4',
+    contextMaxToken: 8000,
+    systemMaxToken: 4000,
+    maxTemperature: 1.2,
+    price: 45
+  }
+};
+const defaultQaModels = {
+  'Gpt35-16k': {
+    model: 'gpt-3.5-turbo',
+    name: 'Gpt35-16k',
+    maxToken: 16000,
+    price: 3
+  }
+};
+const defaultVectorModels = {
+  'text-embedding-ada-002': {
+    model: 'text-embedding-ada-002',
+    name: 'Embedding-2',
+    price: 0.2
+  }
 };
 
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
+  const envs = {
+    beianText: process.env.SAFE_BEIAN_TEXT || '',
+    googleVerKey: process.env.CLIENT_GOOGLE_VER_TOKEN || '',
+    baiduTongji: process.env.BAIDU_TONGJI || ''
+  };
+
   jsonRes<InitDateResponse>(res, {
     data: {
-      beianText: process.env.SAFE_BEIAN_TEXT || '',
-      googleVerKey: process.env.CLIENT_GOOGLE_VER_TOKEN || '',
-      baiduTongji: process.env.BAIDU_TONGJI === '1'
+      ...envs,
+      ...initSystemModels()
     }
   });
 }
+
+export function initSystemModels() {
+  const { chatModels, qaModels, vectorModels } = (() => {
+    try {
+      const chatModels = Object.values(JSON.parse(readFileSync('data/ChatModels.json', 'utf-8')));
+      const qaModels = Object.values(JSON.parse(readFileSync('data/QAModels.json', 'utf-8')));
+      const vectorModels = Object.values(
+        JSON.parse(readFileSync('data/VectorModels.json', 'utf-8'))
+      );
+
+      return {
+        chatModels,
+        qaModels,
+        vectorModels
+      };
+    } catch (error) {
+      console.log(error);
+
+      return {
+        chatModels: Object.values(defaultmodels),
+        qaModels: Object.values(defaultQaModels),
+        vectorModels: Object.values(defaultVectorModels)
+      };
+    }
+  })() as {
+    chatModels: ChatModelItemType[];
+    qaModels: QAModelItemType[];
+    vectorModels: VectorModelItemType[];
+  };
+  global.chatModels = chatModels;
+  global.qaModels = qaModels;
+  global.vectorModels = vectorModels;
+
+  return {
+    chatModels,
+    qaModels,
+    vectorModels
+  };
+}
diff --git a/client/src/pages/api/system/updateEnv.ts b/client/src/pages/api/system/updateEnv.ts
index 1ed408db3..1576cf9bb 100644
--- a/client/src/pages/api/system/updateEnv.ts
+++ b/client/src/pages/api/system/updateEnv.ts
@@ -1,31 +1,22 @@
 import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
-import { System } from '@/service/models/system';
 import { authUser } from '@/service/utils/auth';
-
-export type InitDateResponse = {
-  beianText: string;
-  googleVerKey: string;
-};
+import { readFileSync } from 'fs';
 
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
   await authUser({ req, authRoot: true });
   updateSystemEnv();
-  jsonRes<InitDateResponse>(res);
+  jsonRes(res);
 }
 
 export async function updateSystemEnv() {
   try {
-    const mongoData = await System.findOne();
+    const res = JSON.parse(readFileSync('data/SystemParams.json', 'utf-8'));
 
-    if (mongoData) {
-      const obj = mongoData.toObject();
-      global.systemEnv = {
-        ...global.systemEnv,
-        ...obj
-      };
-    }
-    console.log('update env', global.systemEnv);
+    global.systemEnv = {
+      ...global.systemEnv,
+      ...res
+    };
   } catch (error) {
     console.log('update system env error');
   }
diff --git a/client/src/pages/app/detail/components/edit/components/ChatTest.tsx b/client/src/pages/app/detail/components/edit/components/ChatTest.tsx
index c4197d756..d91af01a5 100644
--- a/client/src/pages/app/detail/components/edit/components/ChatTest.tsx
+++ b/client/src/pages/app/detail/components/edit/components/ChatTest.tsx
@@ -15,6 +15,8 @@ import { SystemInputEnum } from '@/constants/app';
 import { streamFetch } from '@/api/fetch';
 import MyTooltip from '@/components/MyTooltip';
 import ChatBox, { type ComponentRef, type StartChatFnProps } from '@/components/ChatBox';
+import { useToast } from '@/hooks/useToast';
+import { getErrText } from '@/utils/tools';
 
 export type ChatTestComponentRef = {
   resetChatTest: () => void;
@@ -34,6 +36,7 @@ const ChatTest = (
 ) => {
   const BoxRef = useRef(null);
   const ChatBoxRef = useRef<ComponentRef>(null);
+  const { toast } = useToast();
   const isOpen = useMemo(() => modules && modules.length > 0, [modules]);
 
   const variableModules = useMemo(
@@ -60,21 +63,30 @@ const ChatTest = (
       const history = messages.slice(-historyMaxLen - 2, -2);
 
       // 流请求，获取数据
-      const { responseText } = await streamFetch({
+      const { responseText, errMsg } = await streamFetch({
         url: '/api/chat/chatTest',
         data: {
           history,
           prompt: messages[messages.length - 2].content,
           modules,
-          variables
+          variables,
+          appId: app._id,
+          appName: `调试-${app.name}`
         },
         onMessage: generatingMessage,
         abortSignal: controller
       });
 
+      if (errMsg) {
+        return Promise.reject({
+          message: errMsg,
+          responseText
+        });
+      }
+
       return { responseText };
     },
-    [modules]
+    [app._id, app.name, modules]
   );
 
   useOutsideClick({
diff --git a/client/src/pages/app/detail/components/edit/components/NodeCQNode.tsx b/client/src/pages/app/detail/components/edit/components/NodeRINode.tsx
similarity index 96%
rename from client/src/pages/app/detail/components/edit/components/NodeCQNode.tsx
rename to client/src/pages/app/detail/components/edit/components/NodeRINode.tsx
index 154116033..17e999696 100644
--- a/client/src/pages/app/detail/components/edit/components/NodeCQNode.tsx
+++ b/client/src/pages/app/detail/components/edit/components/NodeRINode.tsx
@@ -6,14 +6,14 @@ import { FlowModuleItemType } from '@/types/flow';
 import Divider from './modules/Divider';
 import Container from './modules/Container';
 import RenderInput from './render/RenderInput';
-import type { ClassifyQuestionAgentItemType } from '@/types/app';
+import type { RecognizeIntentionAgentItemType } from '@/types/app';
 import { Handle, Position } from 'reactflow';
 import { customAlphabet } from 'nanoid';
 const nanoid = customAlphabet('abcdefghijklmnopqrstuvwxyz1234567890', 4);
 import MyIcon from '@/components/Icon';
 import { FlowOutputItemTypeEnum } from '@/constants/flow';
 
-const NodeCQNode = ({
+const NodeRINode = ({
   data: { moduleId, inputs, outputs, onChangeNode, ...props }
 }: NodeProps<FlowModuleItemType>) => {
   return (
@@ -30,7 +30,7 @@ const NodeCQNode = ({
               value: agents = []
             }: {
               key: string;
-              value?: ClassifyQuestionAgentItemType[];
+              value?: RecognizeIntentionAgentItemType[];
             }) => (
               <Box>
                 {agents.map((item, i) => (
@@ -133,4 +133,4 @@ const NodeCQNode = ({
     </NodeCard>
   );
 };
-export default React.memo(NodeCQNode);
+export default React.memo(NodeRINode);
diff --git a/client/src/pages/app/detail/components/edit/index.tsx b/client/src/pages/app/detail/components/edit/index.tsx
index d7b1eef52..96f8e8e5d 100644
--- a/client/src/pages/app/detail/components/edit/index.tsx
+++ b/client/src/pages/app/detail/components/edit/index.tsx
@@ -49,7 +49,7 @@ const NodeAnswer = dynamic(() => import('./components/NodeAnswer'), {
 const NodeQuestionInput = dynamic(() => import('./components/NodeQuestionInput'), {
   ssr: false
 });
-const NodeCQNode = dynamic(() => import('./components/NodeCQNode'), {
+const NodeRINode = dynamic(() => import('./components/NodeRINode'), {
   ssr: false
 });
 const NodeUserGuide = dynamic(() => import('./components/NodeUserGuide'), {
@@ -70,7 +70,7 @@ const nodeTypes = {
   [FlowModuleTypeEnum.kbSearchNode]: NodeKbSearch,
   [FlowModuleTypeEnum.tfSwitchNode]: NodeTFSwitch,
   [FlowModuleTypeEnum.answerNode]: NodeAnswer,
-  [FlowModuleTypeEnum.classifyQuestionNode]: NodeCQNode
+  [FlowModuleTypeEnum.recognizeIntention]: NodeRINode
 };
 const edgeTypes = {
   buttonedge: ButtonEdge
diff --git a/client/src/pages/chat/components/QuoteModal.tsx b/client/src/pages/chat/components/QuoteModal.tsx
index feb9cc846..7b65e4f75 100644
--- a/client/src/pages/chat/components/QuoteModal.tsx
+++ b/client/src/pages/chat/components/QuoteModal.tsx
@@ -9,7 +9,6 @@ import {
   Box,
   useTheme
 } from '@chakra-ui/react';
-import { QuoteItemType } from '@/pages/api/openapi/kb/appKbSearch';
 import MyIcon from '@/components/Icon';
 import InputDataModal from '@/pages/kb/components/InputDataModal';
 import { getKbDataItemById } from '@/api/plugins/kb';
diff --git a/client/src/pages/index.tsx b/client/src/pages/index.tsx
index a24a12ce9..262f43f70 100644
--- a/client/src/pages/index.tsx
+++ b/client/src/pages/index.tsx
@@ -4,6 +4,7 @@ import Markdown from '@/components/Markdown';
 import { useMarkdown } from '@/hooks/useMarkdown';
 import { useRouter } from 'next/router';
 import { useGlobalStore } from '@/store/global';
+import { beianText } from '@/store/static';
 
 import styles from './index.module.scss';
 import axios from 'axios';
@@ -13,10 +14,7 @@ const Home = () => {
   const router = useRouter();
   const { inviterId } = router.query as { inviterId: string };
   const { data } = useMarkdown({ url: '/intro.md' });
-  const {
-    isPc,
-    initData: { beianText }
-  } = useGlobalStore();
+  const { isPc } = useGlobalStore();
   const [star, setStar] = useState(1500);
 
   useEffect(() => {
diff --git a/client/src/pages/kb/components/InputDataModal.tsx b/client/src/pages/kb/components/InputDataModal.tsx
index 14c3092ad..b4fb75c09 100644
--- a/client/src/pages/kb/components/InputDataModal.tsx
+++ b/client/src/pages/kb/components/InputDataModal.tsx
@@ -17,6 +17,7 @@ import { useToast } from '@/hooks/useToast';
 import { TrainingModeEnum } from '@/constants/plugin';
 import { getErrText } from '@/utils/tools';
 import MyIcon from '@/components/Icon';
+import { vectorModelList } from '@/store/static';
 
 export type FormData = { dataId?: string; a: string; q: string };
 
@@ -65,6 +66,7 @@ const InputDataModal = ({
         };
         const { insertLen } = await postKbDataFromList({
           kbId,
+          model: vectorModelList[0].model,
           mode: TrainingModeEnum.index,
           data: [data]
         });
diff --git a/client/src/pages/kb/components/SelectFileModal.tsx b/client/src/pages/kb/components/SelectFileModal.tsx
index 421486010..22a05e919 100644
--- a/client/src/pages/kb/components/SelectFileModal.tsx
+++ b/client/src/pages/kb/components/SelectFileModal.tsx
@@ -1,4 +1,4 @@
-import React, { useState, useCallback, useRef } from 'react';
+import React, { useState, useCallback } from 'react';
 import {
   Box,
   Flex,
@@ -22,9 +22,9 @@ import Radio from '@/components/Radio';
 import { splitText_token } from '@/utils/file';
 import { TrainingModeEnum } from '@/constants/plugin';
 import { getErrText } from '@/utils/tools';
-import { ChatModelMap, OpenAiChatEnum, embeddingPrice } from '@/constants/model';
 import { formatPrice } from '@/utils/user';
 import MySlider from '@/components/Slider';
+import { qaModelList, vectorModelList } from '@/store/static';
 
 const fileExtension = '.txt,.doc,.docx,.pdf,.md';
 
@@ -39,12 +39,14 @@ const SelectFileModal = ({
 }) => {
   const [modeMap, setModeMap] = useState({
     [TrainingModeEnum.qa]: {
-      maxLen: 8000,
-      price: ChatModelMap[OpenAiChatEnum.GPT3516k].price
+      model: qaModelList[0].model,
+      maxLen: (qaModelList[0]?.maxToken || 16000) * 0.5,
+      price: qaModelList[0]?.price || 3
     },
     [TrainingModeEnum.index]: {
+      model: vectorModelList[0].model,
       maxLen: 600,
-      price: embeddingPrice
+      price: vectorModelList[0]?.price || 0.2
     }
   });
   const [btnLoading, setBtnLoading] = useState(false);
@@ -111,6 +113,7 @@ const SelectFileModal = ({
     },
     [toast]
   );
+  console.log({ model: modeMap[mode].model });
 
   const { mutate, isLoading: uploading } = useMutation({
     mutationFn: async () => {
@@ -122,6 +125,7 @@ const SelectFileModal = ({
       for (let i = 0; i < splitRes.chunks.length; i += step) {
         const { insertLen } = await postKbDataFromList({
           kbId,
+          model: modeMap[mode].model,
           data: splitRes.chunks
             .slice(i, i + step)
             .map((item) => ({ q: item.value, a: '', source: item.filename })),
@@ -275,8 +279,8 @@ const SelectFileModal = ({
                       setModeMap((state) => ({
                         ...state,
                         [TrainingModeEnum.index]: {
-                          maxLen: val,
-                          price: embeddingPrice
+                          ...modeMap[TrainingModeEnum.index],
+                          maxLen: val
                         }
                       }));
                     }}
diff --git a/client/src/pages/number/components/BillTable.tsx b/client/src/pages/number/components/BillTable.tsx
index be7d78040..dd932d816 100644
--- a/client/src/pages/number/components/BillTable.tsx
+++ b/client/src/pages/number/components/BillTable.tsx
@@ -1,6 +1,6 @@
 import React, { useState } from 'react';
 import { Table, Thead, Tbody, Tr, Th, Td, TableContainer, Flex, Box } from '@chakra-ui/react';
-import { BillTypeMap } from '@/constants/user';
+import { BillSourceMap } from '@/constants/user';
 import { getUserBills } from '@/api/user';
 import type { UserBillType } from '@/types/user';
 import { usePagination } from '@/hooks/usePagination';
@@ -39,10 +39,8 @@ const BillTable = () => {
           <Thead>
             <Tr>
               <Th>时间</Th>
-              <Th>类型</Th>
-              <Th>模型</Th>
-              <Th>内容长度</Th>
-              <Th>Tokens 长度</Th>
+              <Th>来源</Th>
+              <Th>应用名</Th>
               <Th>金额</Th>
             </Tr>
           </Thead>
@@ -50,11 +48,9 @@ const BillTable = () => {
             {bills.map((item) => (
               <Tr key={item.id}>
                 <Td>{dayjs(item.time).format('YYYY/MM/DD HH:mm:ss')}</Td>
-                <Td>{BillTypeMap[item.type] || '-'}</Td>
-                <Td>{item.modelName}</Td>
-                <Td>{item.textLen}</Td>
-                <Td>{item.tokenLen}</Td>
-                <Td>{item.price}元</Td>
+                <Td>{BillSourceMap[item.source]}</Td>
+                <Td>{item.appName || '-'}</Td>
+                <Td>{item.total}元</Td>
               </Tr>
             ))}
           </Tbody>
diff --git a/client/src/service/api/axios.ts b/client/src/service/api/axios.ts
new file mode 100644
index 000000000..764f8c607
--- /dev/null
+++ b/client/src/service/api/axios.ts
@@ -0,0 +1,114 @@
+import axios, { Method, InternalAxiosRequestConfig, AxiosResponse } from 'axios';
+
+interface ConfigType {
+  headers?: { [key: string]: string };
+  hold?: boolean;
+  timeout?: number;
+}
+interface ResponseDataType {
+  code: number;
+  message: string;
+  data: any;
+}
+
+/**
+ * 请求开始
+ */
+function requestStart(config: InternalAxiosRequestConfig): InternalAxiosRequestConfig {
+  if (config.headers) {
+    // config.headers.Authorization = getToken();
+  }
+
+  return config;
+}
+
+/**
+ * 请求成功,检查请求头
+ */
+function responseSuccess(response: AxiosResponse<ResponseDataType>) {
+  return response;
+}
+/**
+ * 响应数据检查
+ */
+function checkRes(data: ResponseDataType) {
+  if (data === undefined) {
+    console.log('error->', data, 'data is empty');
+    return Promise.reject('服务器异常');
+  } else if (data.code < 200 || data.code >= 400) {
+    return Promise.reject(data);
+  }
+  return data.data;
+}
+
+/**
+ * 响应错误
+ */
+function responseError(err: any) {
+  console.log('error->', '请求错误', err);
+
+  if (!err) {
+    return Promise.reject({ message: '未知错误' });
+  }
+  if (typeof err === 'string') {
+    return Promise.reject({ message: err });
+  }
+  return Promise.reject(err);
+}
+
+/* 创建请求实例 */
+const instance = axios.create({
+  timeout: 60000, // 超时时间
+  headers: {
+    'content-type': 'application/json'
+  }
+});
+
+/* 请求拦截 */
+instance.interceptors.request.use(requestStart, (err) => Promise.reject(err));
+/* 响应拦截 */
+instance.interceptors.response.use(responseSuccess, (err) => Promise.reject(err));
+
+function request(url: string, data: any, config: ConfigType, method: Method): any {
+  /* 去空 */
+  for (const key in data) {
+    if (data[key] === null || data[key] === undefined) {
+      delete data[key];
+    }
+  }
+
+  return instance
+    .request({
+      baseURL: `http://localhost:${process.env.PORT || 3000}/api`,
+      url,
+      method,
+      data: ['POST', 'PUT'].includes(method) ? data : null,
+      params: !['POST', 'PUT'].includes(method) ? data : null,
+      ...config // 用户自定义配置，可以覆盖前面的配置
+    })
+    .then((res) => checkRes(res.data))
+    .catch((err) => responseError(err));
+}
+
+/**
+ * api请求方式
+ * @param {String} url
+ * @param {Any} params
+ * @param {Object} config
+ * @returns
+ */
+export function GET<T>(url: string, params = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, params, config, 'GET');
+}
+
+export function POST<T>(url: string, data = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, data, config, 'POST');
+}
+
+export function PUT<T>(url: string, data = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, data, config, 'PUT');
+}
+
+export function DELETE<T>(url: string, data = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, data, config, 'DELETE');
+}
diff --git a/client/src/service/api/request.ts b/client/src/service/api/request.ts
index 2a93f80fd..33cc6ed21 100644
--- a/client/src/service/api/request.ts
+++ b/client/src/service/api/request.ts
@@ -93,6 +93,8 @@ export const moduleFetch = ({ url, data, res }: Props) =>
                 event: sseResponseEventEnum.answer,
                 data: JSON.stringify(data)
               });
+            } else if (item.event === sseResponseEventEnum.error) {
+              return reject(getErrText(data, '流响应错误'));
             }
           });
           read();
diff --git a/client/src/service/events/generateQA.ts b/client/src/service/events/generateQA.ts
index bfbf6c56f..4647703de 100644
--- a/client/src/service/events/generateQA.ts
+++ b/client/src/service/events/generateQA.ts
@@ -1,15 +1,16 @@
 import { TrainingData } from '@/service/mongo';
-import { getApiKey } from '../utils/auth';
 import { OpenAiChatEnum } from '@/constants/model';
 import { pushSplitDataBill } from '@/service/events/pushBill';
 import { openaiAccountError } from '../errorCode';
-import { modelServiceToolMap } from '../utils/chat';
 import { ChatRoleEnum } from '@/constants/chat';
-import { BillTypeEnum } from '@/constants/user';
+import { BillSourceEnum } from '@/constants/user';
 import { pushDataToKb } from '@/pages/api/openapi/kb/pushData';
 import { TrainingModeEnum } from '@/constants/plugin';
 import { ERROR_ENUM } from '../errorCode';
 import { sendInform } from '@/pages/api/user/inform/send';
+import { authBalanceByUid } from '../utils/auth';
+import { axiosConfig, getOpenAIApi } from '../ai/openai';
+import { ChatCompletionRequestMessage } from 'openai';
 
 const reduceQueue = () => {
   global.qaQueueLen = global.qaQueueLen > 0 ? global.qaQueueLen - 1 : 0;
@@ -37,7 +38,8 @@ export async function generateQA(): Promise<any> {
       kbId: 1,
       prompt: 1,
       q: 1,
-      source: 1
+      source: 1,
+      model: 1
     });
 
     // task preemption
@@ -51,54 +53,59 @@ export async function generateQA(): Promise<any> {
     userId = String(data.userId);
     const kbId = String(data.kbId);
 
-    // 余额校验并获取 openapi Key
-    const { systemAuthKey } = await getApiKey({
-      model: OpenAiChatEnum.GPT35,
-      userId,
-      mustPay: true
-    });
+    await authBalanceByUid(userId);
 
     const startTime = Date.now();
 
+    const chatAPI = getOpenAIApi();
+
     // 请求 chatgpt 获取回答
     const response = await Promise.all(
-      [data.q].map((text) =>
-        modelServiceToolMap
-          .chatCompletion({
-            model: OpenAiChatEnum.GPT3516k,
-            apiKey: systemAuthKey,
-            temperature: 0.8,
-            messages: [
-              {
-                obj: ChatRoleEnum.System,
-                value: `你是出题人.
+      [data.q].map((text) => {
+        const messages: ChatCompletionRequestMessage[] = [
+          {
+            role: 'system',
+            content: `你是出题人.
 ${data.prompt || '用户会发送一段长文本'}.
 从中选出 25 个问题和答案. 答案详细完整. 按格式回答: Q1:
 A1:
 Q2:
 A2:
 ...`
-              },
-              {
-                obj: 'Human',
-                value: text
-              }
-            ],
-            stream: false
-          })
-          .then(({ totalTokens, responseText, responseMessages }) => {
-            const result = formatSplitText(responseText); // 格式化后的QA对
+          },
+          {
+            role: 'user',
+            content: text
+          }
+        ];
+        return chatAPI
+          .createChatCompletion(
+            {
+              model: data.model,
+              temperature: 0.8,
+              messages,
+              stream: false
+            },
+            {
+              timeout: 480000,
+              ...axiosConfig()
+            }
+          )
+          .then((res) => {
+            const answer = res.data.choices?.[0].message?.content;
+            const totalTokens = res.data.usage?.total_tokens || 0;
+
+            const result = formatSplitText(answer || ''); // 格式化后的QA对
             console.log(`split result length: `, result.length);
             // 计费
             pushSplitDataBill({
-              isPay: result.length > 0,
               userId: data.userId,
-              type: BillTypeEnum.QA,
-              textLen: responseMessages.map((item) => item.value).join('').length,
-              totalTokens
+              totalTokens,
+              model: data.model,
+              appName: 'QA 拆分'
             });
             return {
-              rawContent: responseText,
+              rawContent: answer,
               result
             };
           })
@@ -106,8 +113,8 @@ A2:
             console.log('QA拆分错误');
             console.log(err.response?.status, err.response?.statusText, err.response?.data);
             return Promise.reject(err);
-          })
-      )
+          });
+      })
     );
 
     const responseList = response.map((item) => item.result).flat();
@@ -120,6 +127,7 @@ A2:
         source: data.source
       })),
       userId,
+      model: global.vectorModels[0].model,
       mode: TrainingModeEnum.index
     });
 
diff --git a/client/src/service/events/generateVector.ts b/client/src/service/events/generateVector.ts
index 99ca7b9b8..1101df973 100644
--- a/client/src/service/events/generateVector.ts
+++ b/client/src/service/events/generateVector.ts
@@ -1,6 +1,6 @@
 import { openaiAccountError } from '../errorCode';
 import { insertKbItem } from '@/service/pg';
-import { openaiEmbedding } from '@/pages/api/openapi/plugin/openaiEmbedding';
+import { getVector } from '@/pages/api/openapi/plugin/vector';
 import { TrainingData } from '../models/trainingData';
 import { ERROR_ENUM } from '../errorCode';
 import { TrainingModeEnum } from '@/constants/plugin';
@@ -33,7 +33,8 @@ export async function generateVector(): Promise<any> {
       kbId: 1,
       q: 1,
       a: 1,
-      source: 1
+      source: 1,
+      model: 1
     });
 
     // task preemption
@@ -55,10 +56,10 @@ export async function generateVector(): Promise<any> {
     ];
 
     // 生成词向量
-    const vectors = await openaiEmbedding({
+    const vectors = await getVector({
+      model: data.model,
       input: dataItems.map((item) => item.q),
-      userId,
-      mustPay: true
+      userId
     });
 
     // 生成结果插入到 pg
diff --git a/client/src/service/events/pushBill.ts b/client/src/service/events/pushBill.ts
index df5432e9f..81ff070f8 100644
--- a/client/src/service/events/pushBill.ts
+++ b/client/src/service/events/pushBill.ts
@@ -1,66 +1,85 @@
 import { connectToDatabase, Bill, User, ShareChat } from '../mongo';
-import {
-  ChatModelMap,
-  OpenAiChatEnum,
-  ChatModelType,
-  embeddingModel,
-  embeddingPrice
-} from '@/constants/model';
-import { BillTypeEnum } from '@/constants/user';
+import { BillSourceEnum } from '@/constants/user';
+import { getModel } from '../utils/data';
+import type { BillListItemType } from '@/types/mongoSchema';
 
-export const pushChatBill = async ({
-  isPay,
-  chatModel,
-  userId,
+export const createTaskBill = async ({
+  appName,
   appId,
-  textLen,
-  tokens,
-  type
+  userId,
+  source
 }: {
-  isPay: boolean;
-  chatModel: ChatModelType;
-  userId: string;
+  appName: string;
   appId: string;
-  textLen: number;
-  tokens: number;
-  type: BillTypeEnum.chat | BillTypeEnum.openapiChat;
+  userId: string;
+  source: `${BillSourceEnum}`;
 }) => {
-  console.log(`chat generate success. text len: ${textLen}. token len: ${tokens}. pay:${isPay}`);
-  if (!isPay) return;
+  const res = await Bill.create({
+    userId,
+    appName,
+    appId,
+    total: 0,
+    source,
+    list: []
+  });
+  return String(res._id);
+};
 
-  let billId = '';
+export const pushTaskBillListItem = async ({
+  billId,
+  moduleName,
+  amount,
+  model,
+  tokenLen
+}: { billId?: string } & BillListItemType) => {
+  if (!billId) return;
+  try {
+    await Bill.findByIdAndUpdate(billId, {
+      $push: {
+        list: {
+          moduleName,
+          amount,
+          model,
+          tokenLen
+        }
+      }
+    });
+  } catch (error) {}
+};
+export const finishTaskBill = async ({ billId }: { billId: string }) => {
+  try {
+    // update bill
+    const res = await Bill.findByIdAndUpdate(billId, [
+      {
+        $set: {
+          total: {
+            $sum: '$list.amount'
+          },
+          time: new Date()
+        }
+      }
+    ]);
+    if (!res) return;
+    const total = res.list.reduce((sum, item) => sum + item.amount, 0) || 0;
+
+    console.log('finish bill:', total);
+
+    // 账号扣费
+    await User.findByIdAndUpdate(res.userId, {
+      $inc: { balance: -total }
+    });
+  } catch (error) {
+    console.log('Finish bill failed:', error);
+    billId && Bill.findByIdAndDelete(billId);
+  }
+};
+
+export const delTaskBill = async (billId?: string) => {
+  if (!billId) return;
 
   try {
-    await connectToDatabase();
-
-    // 计算价格
-    const unitPrice = ChatModelMap[chatModel]?.price || 3;
-    const price = unitPrice * tokens;
-
-    try {
-      // 插入 Bill 记录
-      const res = await Bill.create({
-        userId,
-        type,
-        modelName: chatModel,
-        appId,
-        textLen,
-        tokenLen: tokens,
-        price
-      });
-      billId = res._id;
-
-      // 账号扣费
-      await User.findByIdAndUpdate(userId, {
-        $inc: { balance: -price }
-      });
-    } catch (error) {
-      console.log('创建账单失败:', error);
-      billId && Bill.findByIdAndDelete(billId);
-    }
-  } catch (error) {
-    console.log(error);
-  }
+    await Bill.findByIdAndRemove(billId);
+  } catch (error) {}
 };
 
 export const updateShareChatBill = async ({
@@ -81,22 +100,17 @@ export const updateShareChatBill = async ({
 };
 
 export const pushSplitDataBill = async ({
-  isPay,
   userId,
   totalTokens,
-  textLen,
-  type
+  model,
+  appName
 }: {
-  isPay: boolean;
+  model: string;
   userId: string;
   totalTokens: number;
-  textLen: number;
-  type: BillTypeEnum.QA;
+  appName: string;
 }) => {
-  console.log(
-    `splitData generate success. text len: ${textLen}. token len: ${totalTokens}. pay:${isPay}`
-  );
-  if (!isPay) return;
+  console.log(`splitData generate success. token len: ${totalTokens}.`);
 
   let billId;
 
@@ -104,24 +118,22 @@ export const pushSplitDataBill = async ({
     await connectToDatabase();
 
     // 获取模型单价格, 都是用 gpt35 拆分
-    const unitPrice = ChatModelMap[OpenAiChatEnum.GPT3516k].price || 3;
+    const unitPrice = global.chatModels.find((item) => item.model === model)?.price || 3;
     // 计算价格
-    const price = unitPrice * totalTokens;
+    const total = unitPrice * totalTokens;
 
     // 插入 Bill 记录
     const res = await Bill.create({
       userId,
-      type,
-      modelName: OpenAiChatEnum.GPT3516k,
-      textLen,
+      appName,
       tokenLen: totalTokens,
-      price
+      total
     });
     billId = res._id;
 
     // 账号扣费
     await User.findByIdAndUpdate(userId, {
-      $inc: { balance: -price }
+      $inc: { balance: -total }
     });
   } catch (error) {
     console.log('创建账单失败:', error);
@@ -130,21 +142,14 @@ export const pushSplitDataBill = async ({
 };
 
 export const pushGenerateVectorBill = async ({
-  isPay,
   userId,
-  text,
-  tokenLen
+  tokenLen,
+  model
 }: {
-  isPay: boolean;
   userId: string;
-  text: string;
   tokenLen: number;
+  model: string;
 }) => {
-  // console.log(
-  //   `vector generate success. text len: ${text.length}. token len: ${tokenLen}. pay:${isPay}`
-  // );
-  if (!isPay) return;
-
   let billId;
 
   try {
@@ -152,23 +157,22 @@ export const pushGenerateVectorBill = async ({
 
     try {
       // 计算价格. 至少为1
-      let price = embeddingPrice * tokenLen;
-      price = price > 1 ? price : 1;
+      const unitPrice = global.vectorModels.find((item) => item.model === model)?.price || 0.2;
+      let total = unitPrice * tokenLen;
+      total = total > 1 ? total : 1;
 
       // 插入 Bill 记录
       const res = await Bill.create({
         userId,
-        type: BillTypeEnum.vector,
-        modelName: embeddingModel,
-        textLen: text.length,
-        tokenLen,
-        price
+        model,
+        appName: '索引生成',
+        total
       });
       billId = res._id;
 
       // 账号扣费
       await User.findByIdAndUpdate(userId, {
-        $inc: { balance: -price }
+        $inc: { balance: -total }
       });
     } catch (error) {
       console.log('创建账单失败:', error);
@@ -178,3 +182,9 @@ export const pushGenerateVectorBill = async ({
     console.log(error);
   }
 };
+
+export const countModelPrice = ({ model, tokens }: { model: string; tokens: number }) => {
+  const modelData = getModel(model);
+  if (!modelData) return 0;
+  return modelData.price * tokens;
+};
diff --git a/client/src/service/models/app.ts b/client/src/service/models/app.ts
index 2efffc54c..7a969ef67 100644
--- a/client/src/service/models/app.ts
+++ b/client/src/service/models/app.ts
@@ -1,6 +1,5 @@
 import { Schema, model, models, Model } from 'mongoose';
 import { AppSchema as AppType } from '@/types/mongoSchema';
-import { ChatModelMap, OpenAiChatEnum } from '@/constants/model';
 
 const AppSchema = new Schema({
   userId: {
@@ -24,50 +23,6 @@ const AppSchema = new Schema({
     type: Date,
     default: () => new Date()
   },
-  chat: {
-    relatedKbs: {
-      type: [Schema.Types.ObjectId],
-      ref: 'kb',
-      default: []
-    },
-    searchSimilarity: {
-      type: Number,
-      default: 0.8
-    },
-    searchLimit: {
-      type: Number,
-      default: 5
-    },
-    searchEmptyText: {
-      type: String,
-      default: ''
-    },
-    systemPrompt: {
-      type: String,
-      default: ''
-    },
-    limitPrompt: {
-      type: String,
-      default: ''
-    },
-    maxToken: {
-      type: Number,
-      default: 4000,
-      min: 100
-    },
-    temperature: {
-      type: Number,
-      min: 0,
-      max: 10,
-      default: 0
-    },
-    chatModel: {
-      // 聊天时使用的模型
-      type: String,
-      enum: Object.keys(ChatModelMap),
-      default: OpenAiChatEnum.GPT3516k
-    }
-  },
   share: {
     topNum: {
       type: Number,
diff --git a/client/src/service/models/bill.ts b/client/src/service/models/bill.ts
index e1ce02db8..3f327079c 100644
--- a/client/src/service/models/bill.ts
+++ b/client/src/service/models/bill.ts
@@ -1,7 +1,6 @@
 import { Schema, model, models, Model } from 'mongoose';
-import { ChatModelMap, embeddingModel } from '@/constants/model';
 import { BillSchema as BillType } from '@/types/mongoSchema';
-import { BillTypeMap } from '@/constants/user';
+import { BillSourceEnum, BillSourceMap } from '@/constants/user';
 
 const BillSchema = new Schema({
   userId: {
@@ -9,36 +8,48 @@ const BillSchema = new Schema({
     ref: 'user',
     required: true
   },
-  type: {
+  appName: {
     type: String,
-    enum: Object.keys(BillTypeMap),
-    required: true
-  },
-  modelName: {
-    type: String,
-    enum: [...Object.keys(ChatModelMap), embeddingModel]
+    default: ''
   },
   appId: {
     type: Schema.Types.ObjectId,
-    ref: 'app'
+    ref: 'app',
+    required: false
   },
   time: {
     type: Date,
     default: () => new Date()
   },
-  textLen: {
-    // 提示词+响应的总字数
+  total: {
     type: Number,
     required: true
   },
-  tokenLen: {
-    // 折算成 token 的数量
-    type: Number,
-    required: true
+  source: {
+    type: String,
+    enum: Object.keys(BillSourceMap),
+    default: BillSourceEnum.fastgpt
   },
-  price: {
-    type: Number,
-    required: true
+  list: {
+    type: [
+      {
+        moduleName: {
+          type: String,
+          required: true
+        },
+        amount: {
+          type: Number,
+          required: true
+        },
+        model: {
+          type: String
+        },
+        tokenLen: {
+          type: Number
+        }
+      }
+    ],
+    default: []
   }
 });
 
diff --git a/client/src/service/models/system.ts b/client/src/service/models/system.ts
deleted file mode 100644
index 5762eb5cc..000000000
--- a/client/src/service/models/system.ts
+++ /dev/null
@@ -1,22 +0,0 @@
-import { Schema, model, models } from 'mongoose';
-
-const SystemSchema = new Schema({
-  vectorMaxProcess: {
-    type: Number,
-    default: 10
-  },
-  qaMaxProcess: {
-    type: Number,
-    default: 10
-  },
-  pgIvfflatProbe: {
-    type: Number,
-    default: 10
-  },
-  sensitiveCheck: {
-    type: Boolean,
-    default: false
-  }
-});
-
-export const System = models['system'] || model('system', SystemSchema);
diff --git a/client/src/service/models/trainingData.ts b/client/src/service/models/trainingData.ts
index 65fd96c78..476ee1505 100644
--- a/client/src/service/models/trainingData.ts
+++ b/client/src/service/models/trainingData.ts
@@ -28,13 +28,16 @@ const TrainingDataSchema = new Schema({
     enum: Object.keys(TrainingTypeMap),
     required: true
   },
+  model: {
+    type: String,
+    required: true
+  },
   prompt: {
-    // 拆分时的提示词
+    // qa split prompt
     type: String,
     default: ''
   },
   q: {
-    // 如果是
     type: String,
     default: ''
   },
diff --git a/client/src/service/mongo.ts b/client/src/service/mongo.ts
index 14cb36ae6..7805be576 100644
--- a/client/src/service/mongo.ts
+++ b/client/src/service/mongo.ts
@@ -2,6 +2,7 @@ import mongoose from 'mongoose';
 import tunnel from 'tunnel';
 import { startQueue } from './utils/tools';
 import { updateSystemEnv } from '@/pages/api/system/updateEnv';
+import { initSystemModels } from '@/pages/api/system/getInitData';
 
 /**
  * 连接 MongoDB 数据库
@@ -10,6 +11,7 @@ export async function connectToDatabase(): Promise<void> {
   if (global.mongodb) {
     return;
   }
+  global.mongodb = 'connecting';
 
   // init global data
   global.qaQueueLen = 0;
@@ -31,8 +33,9 @@ export async function connectToDatabase(): Promise<void> {
       }
     });
   }
+  initSystemModels();
+  updateSystemEnv();
 
-  global.mongodb = 'connecting';
   try {
     mongoose.set('strictQuery', true);
     global.mongodb = await mongoose.connect(process.env.MONGODB_URI as string, {
@@ -49,7 +52,6 @@ export async function connectToDatabase(): Promise<void> {
   }
 
   // init function
-  updateSystemEnv();
   startQueue();
 }
 
@@ -66,5 +68,4 @@ export * from './models/collection';
 export * from './models/shareChat';
 export * from './models/kb';
 export * from './models/inform';
-export * from './models/system';
 export * from './models/image';
diff --git a/client/src/service/response.ts b/client/src/service/response.ts
index 9292f4296..576486f9d 100644
--- a/client/src/service/response.ts
+++ b/client/src/service/response.ts
@@ -92,7 +92,7 @@ export const sseErrRes = (res: NextApiResponse, error: any) => {
   } else if (openaiError[error?.response?.statusText]) {
     msg = openaiError[error.response.statusText];
   }
-  console.log('sse error', error);
+  console.log('sse error => ', error);
 
   sseResponse({
     res,
diff --git a/client/src/service/utils/auth.ts b/client/src/service/utils/auth.ts
index 9d6343856..a44ef2d9f 100644
--- a/client/src/service/utils/auth.ts
+++ b/client/src/service/utils/auth.ts
@@ -1,15 +1,11 @@
 import type { NextApiRequest } from 'next';
 import jwt from 'jsonwebtoken';
 import Cookie from 'cookie';
-import { Chat, App, OpenApi, User, ShareChat, KB } from '../mongo';
+import { App, OpenApi, User, ShareChat, KB } from '../mongo';
 import type { AppSchema } from '@/types/mongoSchema';
-import type { ChatItemType } from '@/types/chat';
-import mongoose from 'mongoose';
 import { defaultApp } from '@/constants/model';
 import { formatPrice } from '@/utils/user';
 import { ERROR_ENUM } from '../errorCode';
-import { ChatModelType, OpenAiChatEnum } from '@/constants/model';
-import { hashPassword } from '@/service/utils/tools';
 
 export type AuthType = 'token' | 'root' | 'apikey';
 
@@ -35,6 +31,19 @@ export const parseCookie = (cookie?: string): Promise<string> => {
   });
 };
 
+/* auth balance */
+export const authBalanceByUid = async (uid: string) => {
+  const user = await User.findById(uid);
+  if (!user) {
+    return Promise.reject(ERROR_ENUM.unAuthorization);
+  }
+
+  if (!user.openaiKey && formatPrice(user.balance) <= 0) {
+    return Promise.reject(ERROR_ENUM.insufficientQuota);
+  }
+  return user;
+};
+
 /* uniform auth user */
 export const authUser = async ({
   req,
@@ -144,14 +153,7 @@ export const authUser = async ({
 
   // balance check
   if (authBalance) {
-    const user = await User.findById(uid);
-    if (!user) {
-      return Promise.reject(ERROR_ENUM.unAuthorization);
-    }
-
-    if (!user.openaiKey && formatPrice(user.balance) <= 0) {
-      return Promise.reject(ERROR_ENUM.insufficientQuota);
-    }
+    await authBalanceByUid(uid);
   }
 
   return {
@@ -166,43 +168,6 @@ export const getSystemOpenAiKey = () => {
   return process.env.ONEAPI_KEY || process.env.OPENAIKEY || '';
 };
 
-/* 获取 api 请求的 key */
-export const getApiKey = async ({
-  model,
-  userId,
-  mustPay = false
-}: {
-  model: ChatModelType;
-  userId: string;
-  mustPay?: boolean;
-}) => {
-  const user = await User.findById(userId, 'openaiKey balance');
-  if (!user) {
-    return Promise.reject(ERROR_ENUM.unAuthorization);
-  }
-
-  const userOpenAiKey = user.openaiKey || '';
-  const systemAuthKey = getSystemOpenAiKey();
-
-  // 有自己的key
-  if (!mustPay && userOpenAiKey) {
-    return {
-      userOpenAiKey,
-      systemAuthKey: ''
-    };
-  }
-
-  // 平台账号余额校验
-  if (formatPrice(user.balance) <= 0) {
-    return Promise.reject(ERROR_ENUM.insufficientQuota);
-  }
-
-  return {
-    userOpenAiKey: '',
-    systemAuthKey
-  };
-};
-
 // 模型使用权校验
 export const authApp = async ({
   appId,
@@ -232,14 +197,6 @@ export const authApp = async ({
     if (userId !== String(app.userId)) return Promise.reject(ERROR_ENUM.unAuthModel);
   }
 
-  // do not share detail info
-  if (!reserveDetail && !app.share.isShareDetail && userId !== String(app.userId)) {
-    app.chat = {
-      ...defaultApp.chat,
-      chatModel: app.chat.chatModel
-    };
-  }
-
   return {
     app,
     showModelDetail: userId === String(app.userId)
diff --git a/client/src/service/utils/chat/index.ts b/client/src/service/utils/chat/index.ts
index f9d2c2292..04f262cae 100644
--- a/client/src/service/utils/chat/index.ts
+++ b/client/src/service/utils/chat/index.ts
@@ -1,13 +1,8 @@
 import { ChatItemType } from '@/types/chat';
 import { modelToolMap } from '@/utils/plugin';
-import type { ChatModelType } from '@/constants/model';
-import { ChatRoleEnum, sseResponseEventEnum } from '@/constants/chat';
-import { sseResponse } from '../tools';
+import { ChatRoleEnum } from '@/constants/chat';
 import { OpenAiChatEnum } from '@/constants/model';
-import { chatResponse, openAiStreamResponse } from './openai';
 import type { NextApiResponse } from 'next';
-import { textAdaptGptResponse } from '@/utils/adapt';
-import { parseStreamChunk } from '@/utils/adapt';
 
 export type ChatCompletionType = {
   apiKey: string;
@@ -36,11 +31,6 @@ export type StreamResponseReturnType = {
   finishMessages: ChatItemType[];
 };
 
-export const modelServiceToolMap = {
-  chatCompletion: chatResponse,
-  streamResponse: openAiStreamResponse
-};
-
 /* delete invalid symbol */
 const simplifyStr = (str = '') =>
   str
@@ -54,7 +44,7 @@ export const ChatContextFilter = ({
   prompts,
   maxTokens
 }: {
-  model: ChatModelType;
+  model: string;
   prompts: ChatItemType[];
   maxTokens: number;
 }) => {
@@ -111,126 +101,3 @@ export const ChatContextFilter = ({
 
   return [...systemPrompts, ...chats];
 };
-
-/* stream response */
-export const resStreamResponse = async ({
-  model,
-  res,
-  chatResponse,
-  prompts
-}: StreamResponseType & {
-  model: ChatModelType;
-}) => {
-  // 创建响应流
-  res.setHeader('Content-Type', 'text/event-stream;charset=utf-8');
-  res.setHeader('Access-Control-Allow-Origin', '*');
-  res.setHeader('X-Accel-Buffering', 'no');
-  res.setHeader('Cache-Control', 'no-cache, no-transform');
-
-  const { responseContent, totalTokens, finishMessages } = await modelServiceToolMap.streamResponse(
-    {
-      chatResponse,
-      prompts,
-      res,
-      model
-    }
-  );
-
-  return { responseContent, totalTokens, finishMessages };
-};
-
-/* stream response */
-export const V2_StreamResponse = async ({
-  model,
-  res,
-  chatResponse,
-  prompts
-}: StreamResponseType & {
-  model: ChatModelType;
-}) => {
-  let responseContent = '';
-  let error: any = null;
-  let truncateData = '';
-  const clientRes = async (data: string) => {
-    //部分代理会导致流式传输时的数据被截断，不为json格式，这里做一个兼容
-    const { content = '' } = (() => {
-      try {
-        if (truncateData) {
-          try {
-            //判断是否为json，如果是的话直接跳过后续拼装操作，注意极端情况下可能出现截断成3截以上情况也可以兼容
-            JSON.parse(data);
-          } catch (e) {
-            data = truncateData + data;
-          }
-          truncateData = '';
-        }
-        const json = JSON.parse(data);
-        const content: string = json?.choices?.[0].delta.content || '';
-        error = json.error;
-        responseContent += content;
-        return { content };
-      } catch (error) {
-        truncateData = data;
-        return {};
-      }
-    })();
-
-    if (res.closed || error) return;
-
-    if (data === '[DONE]') {
-      sseResponse({
-        res,
-        event: sseResponseEventEnum.answer,
-        data: textAdaptGptResponse({
-          text: null,
-          finish_reason: 'stop'
-        })
-      });
-      sseResponse({
-        res,
-        event: sseResponseEventEnum.answer,
-        data: '[DONE]'
-      });
-    } else {
-      sseResponse({
-        res,
-        event: sseResponseEventEnum.answer,
-        data: textAdaptGptResponse({
-          text: content
-        })
-      });
-    }
-  };
-
-  try {
-    for await (const chunk of chatResponse.data as any) {
-      if (res.closed) break;
-      const parse = parseStreamChunk(chunk);
-      parse.forEach((item) => clientRes(item.data));
-    }
-  } catch (error) {
-    console.log('pipe error', error);
-  }
-
-  if (error) {
-    console.log(error);
-    return Promise.reject(error);
-  }
-
-  // count tokens
-  const finishMessages = prompts.concat({
-    obj: ChatRoleEnum.AI,
-    value: responseContent
-  });
-
-  const totalTokens = modelToolMap.countTokens({
-    model,
-    messages: finishMessages
-  });
-
-  return {
-    responseContent,
-    totalTokens,
-    finishMessages
-  };
-};
diff --git a/client/src/service/utils/chat/openai.ts b/client/src/service/utils/chat/openai.ts
deleted file mode 100644
index 30cb96184..000000000
--- a/client/src/service/utils/chat/openai.ts
+++ /dev/null
@@ -1,133 +0,0 @@
-import { Configuration, OpenAIApi } from 'openai';
-import { axiosConfig } from '../tools';
-import { ChatModelMap, OpenAiChatEnum } from '@/constants/model';
-import { adaptChatItem_openAI } from '@/utils/plugin/openai';
-import { modelToolMap } from '@/utils/plugin';
-import { ChatCompletionType, ChatContextFilter, StreamResponseType } from './index';
-import { ChatRoleEnum } from '@/constants/chat';
-import { parseStreamChunk } from '@/utils/adapt';
-
-export const getOpenAIApi = (apiKey: string) => {
-  const openaiBaseUrl = process.env.OPENAI_BASE_URL || 'https://api.openai.com/v1';
-  return new OpenAIApi(
-    new Configuration({
-      basePath: apiKey === process.env.ONEAPI_KEY ? process.env.ONEAPI_URL : openaiBaseUrl
-    })
-  );
-};
-
-/* 模型对话 */
-export const chatResponse = async ({
-  model,
-  apiKey,
-  temperature,
-  maxToken = 4000,
-  messages,
-  stream
-}: ChatCompletionType & { model: `${OpenAiChatEnum}` }) => {
-  const modelTokenLimit = ChatModelMap[model]?.contextMaxToken || 4000;
-  const filterMessages = ChatContextFilter({
-    model,
-    prompts: messages,
-    maxTokens: Math.ceil(modelTokenLimit - 300) // filter token. not response maxToken
-  });
-
-  const adaptMessages = adaptChatItem_openAI({ messages: filterMessages, reserveId: false });
-  const chatAPI = getOpenAIApi(apiKey);
-
-  const promptsToken = modelToolMap.countTokens({
-    model,
-    messages: filterMessages
-  });
-
-  maxToken = maxToken + promptsToken > modelTokenLimit ? modelTokenLimit - promptsToken : maxToken;
-
-  const response = await chatAPI.createChatCompletion(
-    {
-      model,
-      temperature: Number(temperature || 0),
-      max_tokens: maxToken,
-      messages: adaptMessages,
-      frequency_penalty: 0.5, // 越大，重复内容越少
-      presence_penalty: -0.5, // 越大，越容易出现新内容
-      stream
-      // stop: ['.!?。']
-    },
-    {
-      timeout: stream ? 60000 : 480000,
-      responseType: stream ? 'stream' : 'json',
-      ...axiosConfig(apiKey)
-    }
-  );
-
-  const responseText = stream ? '' : response.data.choices?.[0].message?.content || '';
-  const totalTokens = stream ? 0 : response.data.usage?.total_tokens || 0;
-
-  return {
-    streamResponse: response,
-    responseMessages: filterMessages.concat({ obj: 'AI', value: responseText }),
-    responseText,
-    totalTokens
-  };
-};
-
-/* openai stream response */
-export const openAiStreamResponse = async ({
-  res,
-  model,
-  chatResponse,
-  prompts
-}: StreamResponseType & {
-  model: `${OpenAiChatEnum}`;
-}) => {
-  try {
-    let responseContent = '';
-
-    const clientRes = async (data: string) => {
-      const { content = '' } = (() => {
-        try {
-          const json = JSON.parse(data);
-          const content: string = json?.choices?.[0].delta.content || '';
-          responseContent += content;
-          return { content };
-        } catch (error) {
-          return {};
-        }
-      })();
-
-      if (data === '[DONE]') return;
-
-      !res.closed && content && res.write(content);
-    };
-
-    try {
-      for await (const chunk of chatResponse.data as any) {
-        if (res.closed) break;
-
-        const parse = parseStreamChunk(chunk);
-        parse.forEach((item) => clientRes(item.data));
-      }
-    } catch (error) {
-      console.log('pipe error', error);
-    }
-
-    // count tokens
-    const finishMessages = prompts.concat({
-      obj: ChatRoleEnum.AI,
-      value: responseContent
-    });
-
-    const totalTokens = modelToolMap.countTokens({
-      model,
-      messages: finishMessages
-    });
-
-    return {
-      responseContent,
-      totalTokens,
-      finishMessages
-    };
-  } catch (error) {
-    return Promise.reject(error);
-  }
-};
diff --git a/client/src/service/utils/data.ts b/client/src/service/utils/data.ts
new file mode 100644
index 000000000..b35ef6817
--- /dev/null
+++ b/client/src/service/utils/data.ts
@@ -0,0 +1,14 @@
+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);
+};
+export const getQAModel = (model: string) => {
+  return global.qaModels.find((item) => item.model === model);
+};
+export const getModel = (model: string) => {
+  return [...global.chatModels, ...global.vectorModels, ...global.qaModels].find(
+    (item) => item.model === model
+  );
+};
diff --git a/client/src/service/utils/tools.ts b/client/src/service/utils/tools.ts
index f0589b6f5..ced563159 100644
--- a/client/src/service/utils/tools.ts
+++ b/client/src/service/utils/tools.ts
@@ -4,7 +4,6 @@ import crypto from 'crypto';
 import jwt from 'jsonwebtoken';
 import { generateQA } from '../events/generateQA';
 import { generateVector } from '../events/generateVector';
-import { sseResponseEventEnum } from '@/constants/chat';
 
 /* 密码加密 */
 export const hashPassword = (psw: string) => {
@@ -33,20 +32,6 @@ export const clearCookie = (res: NextApiResponse) => {
   res.setHeader('Set-Cookie', 'token=; Path=/; Max-Age=0');
 };
 
-/* openai axios config */
-export const axiosConfig = (apikey: string) => {
-  const openaiBaseUrl = process.env.OPENAI_BASE_URL || 'https://api.openai.com/v1';
-
-  return {
-    baseURL: apikey === process.env.ONEAPI_KEY ? process.env.ONEAPI_URL : openaiBaseUrl, // 此处仅对非 npm 模块有效
-    httpsAgent: global.httpsAgent,
-    headers: {
-      Authorization: `Bearer ${apikey}`,
-      auth: process.env.OPENAI_BASE_URL_AUTH || ''
-    }
-  };
-};
-
 export function withNextCors(handler: NextApiHandler): NextApiHandler {
   return async function nextApiHandlerWrappedWithNextCors(
     req: NextApiRequest,
diff --git a/client/src/store/global.ts b/client/src/store/global.ts
index ac0dec08e..916908455 100644
--- a/client/src/store/global.ts
+++ b/client/src/store/global.ts
@@ -1,12 +1,8 @@
 import { create } from 'zustand';
 import { devtools } from 'zustand/middleware';
 import { immer } from 'zustand/middleware/immer';
-import type { InitDateResponse } from '@/pages/api/system/getInitData';
-import { getInitData } from '@/api/system';
 
 type State = {
-  initData: InitDateResponse;
-  loadInitData: () => Promise<void>;
   loading: boolean;
   setLoading: (val: boolean) => null;
   screenWidth: number;
@@ -17,19 +13,6 @@ type State = {
 export const useGlobalStore = create<State>()(
   devtools(
     immer((set, get) => ({
-      initData: {
-        beianText: '',
-        googleVerKey: '',
-        baiduTongji: false
-      },
-      async loadInitData() {
-        try {
-          const res = await getInitData();
-          set((state) => {
-            state.initData = res;
-          });
-        } catch (error) {}
-      },
       loading: false,
       setLoading: (val: boolean) => {
         set((state) => {
diff --git a/client/src/store/static.ts b/client/src/store/static.ts
new file mode 100644
index 000000000..ffdbbf64b
--- /dev/null
+++ b/client/src/store/static.ts
@@ -0,0 +1,36 @@
+import {
+  type QAModelItemType,
+  type ChatModelItemType,
+  type VectorModelItemType
+} from '@/types/model';
+import type { InitDateResponse } from '@/pages/api/system/getInitData';
+import { getInitData } from '@/api/system';
+import { delay } from '@/utils/tools';
+
+export let beianText = '';
+export let googleVerKey = '';
+export let baiduTongji = '';
+export let chatModelList: ChatModelItemType[] = [];
+export let qaModelList: QAModelItemType[] = [];
+export let vectorModelList: VectorModelItemType[] = [];
+
+let retryTimes = 3;
+
+export const clientInitData = async (): Promise<InitDateResponse> => {
+  try {
+    const res = await getInitData();
+
+    chatModelList = res.chatModels;
+    qaModelList = res.qaModels;
+    vectorModelList = res.vectorModels;
+    beianText = res.beianText;
+    googleVerKey = res.googleVerKey;
+    baiduTongji = res.baiduTongji;
+
+    return res;
+  } catch (error) {
+    retryTimes--;
+    await delay(500);
+    return clientInitData();
+  }
+};
diff --git a/client/src/types/app.d.ts b/client/src/types/app.d.ts
index f978be00a..d9c532e35 100644
--- a/client/src/types/app.d.ts
+++ b/client/src/types/app.d.ts
@@ -42,7 +42,7 @@ export type ShareChatEditType = {
 
 /* agent */
 /* question classify */
-export type ClassifyQuestionAgentItemType = {
+export type RecognizeIntentionAgentItemType = {
   value: string;
   key: string;
 };
diff --git a/client/src/types/chat.d.ts b/client/src/types/chat.d.ts
index a8f93486d..4902a0968 100644
--- a/client/src/types/chat.d.ts
+++ b/client/src/types/chat.d.ts
@@ -8,9 +8,6 @@ export type ChatItemType = {
   _id?: string;
   obj: `${ChatRoleEnum}`;
   value: string;
-  quoteLen?: number;
-  quote?: QuoteItemType[];
-  systemPrompt?: string;
   [key: string]: any;
 };
 
diff --git a/client/src/types/index.d.ts b/client/src/types/index.d.ts
index ac15af282..ac226047f 100644
--- a/client/src/types/index.d.ts
+++ b/client/src/types/index.d.ts
@@ -2,6 +2,7 @@ import type { Mongoose } from 'mongoose';
 import type { Agent } from 'http';
 import type { Pool } from 'pg';
 import type { Tiktoken } from '@dqbd/tiktoken';
+import { ChatModelItemType, QAModelItemType, VectorModelItemType } from './model';
 
 export type PagingData<T> = {
   pageNum: number;
@@ -16,9 +17,6 @@ declare global {
   var mongodb: Mongoose | string | null;
   var pgClient: Pool | null;
   var httpsAgent: Agent;
-  var particlesJS: any;
-  var grecaptcha: any;
-  var QRCode: any;
   var qaQueueLen: number;
   var vectorQueueLen: number;
   var OpenAiEncMap: Tiktoken;
@@ -30,8 +28,14 @@ declare global {
     pgIvfflatProbe: number;
     sensitiveCheck: boolean;
   };
+  var chatModels: ChatModelItemType[];
+  var qaModels: QAModelItemType[];
+  var vectorModels: VectorModelItemType[];
 
   interface Window {
     ['pdfjs-dist/build/pdf']: any;
+    particlesJS: any;
+    grecaptcha: any;
+    QRCode: any;
   }
 }
diff --git a/client/src/types/model.d.ts b/client/src/types/model.d.ts
new file mode 100644
index 000000000..715457bf5
--- /dev/null
+++ b/client/src/types/model.d.ts
@@ -0,0 +1,19 @@
+export type ChatModelItemType = {
+  model: string;
+  name: string;
+  contextMaxToken: number;
+  systemMaxToken: number;
+  maxTemperature: number;
+  price: number;
+};
+export type QAModelItemType = {
+  model: string;
+  name: string;
+  maxToken: number;
+  price: number;
+};
+export type VectorModelItemType = {
+  model: string;
+  name: string;
+  price: number;
+};
diff --git a/client/src/types/mongoSchema.d.ts b/client/src/types/mongoSchema.d.ts
index f1d9b4b2b..ac8f0e6ae 100644
--- a/client/src/types/mongoSchema.d.ts
+++ b/client/src/types/mongoSchema.d.ts
@@ -1,7 +1,7 @@
 import type { ChatItemType } from './chat';
 import { ModelNameEnum, ChatModelType, EmbeddingModelType } from '@/constants/model';
 import type { DataType } from './data';
-import { BillTypeEnum, InformTypeEnum } from '@/constants/user';
+import { BillSourceEnum, InformTypeEnum } from '@/constants/user';
 import { TrainingModeEnum } from '@/constants/plugin';
 import type { AppModuleItemType } from './app';
 
@@ -38,17 +38,6 @@ export interface AppSchema {
   avatar: string;
   intro: string;
   updateTime: number;
-  chat: {
-    relatedKbs: string[];
-    searchSimilarity: number;
-    searchLimit: number;
-    searchEmptyText: string;
-    systemPrompt: string;
-    limitPrompt: string;
-    temperature: number;
-    maxToken: number;
-    chatModel: ChatModelType; // 聊天时用的模型，训练后就是训练的模型
-  };
   share: {
     isShare: boolean;
     isShareDetail: boolean;
@@ -68,6 +57,7 @@ export interface TrainingDataSchema {
   kbId: string;
   expireAt: Date;
   lockTime: Date;
+  model: string;
   mode: `${TrainingModeEnum}`;
   prompt: string;
   q: string;
@@ -87,16 +77,21 @@ export interface ChatSchema {
   content: ChatItemType[];
 }
 
+export type BillListItemType = {
+  moduleName: string;
+  amount: number;
+  model?: string;
+  tokenLen?: number;
+};
 export interface BillSchema {
   _id: string;
   userId: string;
-  type: `${BillTypeEnum}`;
-  modelName?: ChatModelType | EmbeddingModelType;
+  appName: string;
   appId?: string;
+  source: `${BillSourceEnum}`;
   time: Date;
-  textLen: number;
-  tokenLen: number;
-  price: number;
+  total: number;
+  list: BillListItemType[];
 }
 
 export interface PaySchema {
diff --git a/client/src/types/user.d.ts b/client/src/types/user.d.ts
index 421ab7734..ca0433938 100644
--- a/client/src/types/user.d.ts
+++ b/client/src/types/user.d.ts
@@ -1,3 +1,4 @@
+import { BillSourceEnum } from '@/constants/user';
 import type { BillSchema } from './mongoSchema';
 export interface UserType {
   _id: string;
@@ -19,9 +20,7 @@ export interface UserUpdateParams {
 export interface UserBillType {
   id: string;
   time: Date;
-  modelName: string;
-  type: BillSchema['type'];
-  textLen: number;
-  tokenLen: number;
-  price: number;
+  appName: string;
+  source: BillSchema['source'];
+  total: number;
 }
diff --git a/client/src/utils/adapt.ts b/client/src/utils/adapt.ts
index b0ae8d658..d6dc4ff21 100644
--- a/client/src/utils/adapt.ts
+++ b/client/src/utils/adapt.ts
@@ -5,7 +5,6 @@ import { ChatItemType } from '@/types/chat';
 import { ChatCompletionRequestMessageRoleEnum } from 'openai';
 import { ChatRoleEnum } from '@/constants/chat';
 import type { MessageItemType } from '@/pages/api/openapi/v1/chat/completions';
-import { ChatModelMap, OpenAiChatEnum } from '@/constants/model';
 import type { AppModuleItemType } from '@/types/app';
 import type { FlowModuleItemType } from '@/types/flow';
 import type { Edge, Node } from 'reactflow';
@@ -16,12 +15,10 @@ const nanoid = customAlphabet('abcdefghijklmnopqrstuvwxyz1234567890', 6);
 export const adaptBill = (bill: BillSchema): UserBillType => {
   return {
     id: bill._id,
-    type: bill.type,
-    modelName: ChatModelMap[bill.modelName as `${OpenAiChatEnum}`]?.name || bill.modelName,
+    source: bill.source,
     time: bill.time,
-    textLen: bill.textLen,
-    tokenLen: bill.tokenLen,
-    price: formatPrice(bill.price)
+    total: formatPrice(bill.total),
+    appName: bill.appName
   };
 };
 
diff --git a/client/src/utils/plugin/openai.ts b/client/src/utils/plugin/openai.ts
index cf2595469..878e6a9a0 100644
--- a/client/src/utils/plugin/openai.ts
+++ b/client/src/utils/plugin/openai.ts
@@ -50,10 +50,10 @@ export const adaptChatItem_openAI = ({
 
 export function countOpenAIToken({
   messages,
-  model
+  model = 'gpt-3.5-turbo'
 }: {
   messages: ChatItemType[];
-  model: `${OpenAiChatEnum}`;
+  model?: string;
 }) {
   const diffVal = model.startsWith('gpt-3.5-turbo') ? 3 : 2;
 
@@ -74,7 +74,7 @@ export const openAiSliceTextByToken = ({
   text,
   length
 }: {
-  model: `${OpenAiChatEnum}`;
+  model: string;
   text: string;
   length: number;
 }) => {