merge main

README.md

@@ -63,7 +63,7 @@ For enterprise inquiries, please contact: **business@nextchat.dev**
 
 企业版咨询: **business@nextchat.dev**
 
-<img width="300" src="https://github.com/user-attachments/assets/3daeb7b6-ab63-4542-9141-2e4a12c80601">
+<img width="300" src="https://github.com/user-attachments/assets/3d4305ac-6e95-489e-884b-51d51db5f692">
 
 ## Features
 
@@ -100,6 +100,7 @@ For enterprise inquiries, please contact: **business@nextchat.dev**
 
 ## What's New
 
+- 🚀 v2.15.4 The Application supports using Tauri fetch LLM API, MORE SECURITY! [#5379](https://github.com/ChatGPTNextWeb/ChatGPT-Next-Web/issues/5379)
 - 🚀 v2.15.0 Now supports Plugins! Read this: [NextChat-Awesome-Plugins](https://github.com/ChatGPTNextWeb/NextChat-Awesome-Plugins)
 - 🚀 v2.14.0 Now supports  Artifacts & SD 
 - 🚀 v2.10.1 support Google Gemini Pro model.
@@ -137,6 +138,7 @@ For enterprise inquiries, please contact: **business@nextchat.dev**
 
 ## 最新动态
 
+- 🚀 v2.15.4 客户端支持Tauri本地直接调用大模型API，更安全！[#5379](https://github.com/ChatGPTNextWeb/ChatGPT-Next-Web/issues/5379)
 - 🚀 v2.15.0 现在支持插件功能了！了解更多：[NextChat-Awesome-Plugins](https://github.com/ChatGPTNextWeb/NextChat-Awesome-Plugins)
 - 🚀 v2.14.0 现在支持 Artifacts & SD 了。
 - 🚀 v2.10.1 现在支持 Gemini Pro 模型。
@@ -332,9 +334,9 @@ To control custom models, use `+` to add a custom model, use `-` to hide a model
 
 User `-all` to disable all default models, `+all` to enable all default models.
 
-For Azure: use `modelName@azure=deploymentName` to customize model name and deployment name.
+For Azure: use `modelName@Azure=deploymentName` to customize model name and deployment name.
-> Example: `+gpt-3.5-turbo@azure=gpt35` will show option `gpt35(Azure)` in model list.
+> Example: `+gpt-3.5-turbo@Azure=gpt35` will show option `gpt35(Azure)` in model list.
-> If you only can use Azure model, `-all,+gpt-3.5-turbo@azure=gpt35` will `gpt35(Azure)` the only option in model list.
+> If you only can use Azure model, `-all,+gpt-3.5-turbo@Azure=gpt35` will `gpt35(Azure)` the only option in model list.
 
 For ByteDance: use `modelName@bytedance=deploymentName` to customize model name and deployment name.
 > Example: `+Doubao-lite-4k@bytedance=ep-xxxxx-xxx` will show option `Doubao-lite-4k(ByteDance)` in model list.

README_CN.md

@@ -216,9 +216,9 @@ ByteDance Api Url.
 
 用来控制模型列表，使用 `+` 增加一个模型，使用 `-` 来隐藏一个模型，使用 `模型名=展示名` 来自定义模型的展示名，用英文逗号隔开。
 
-在Azure的模式下，支持使用`modelName@azure=deploymentName`的方式配置模型名称和部署名称(deploy-name)
+在Azure的模式下，支持使用`modelName@Azure=deploymentName`的方式配置模型名称和部署名称(deploy-name)
-> 示例：`+gpt-3.5-turbo@azure=gpt35`这个配置会在模型列表显示一个`gpt35(Azure)`的选项。
+> 示例：`+gpt-3.5-turbo@Azure=gpt35`这个配置会在模型列表显示一个`gpt35(Azure)`的选项。
-> 如果你只能使用Azure模式，那么设置 `-all,+gpt-3.5-turbo@azure=gpt35` 则可以让对话的默认使用 `gpt35(Azure)`
+> 如果你只能使用Azure模式，那么设置 `-all,+gpt-3.5-turbo@Azure=gpt35` 则可以让对话的默认使用 `gpt35(Azure)`
 
 在ByteDance的模式下，支持使用`modelName@bytedance=deploymentName`的方式配置模型名称和部署名称(deploy-name)
 > 示例: `+Doubao-lite-4k@bytedance=ep-xxxxx-xxx`这个配置会在模型列表显示一个`Doubao-lite-4k(ByteDance)`的选项

README_JA.md

@@ -207,8 +207,8 @@ ByteDance API の URL。
 
 モデルリストを管理します。`+` でモデルを追加し、`-` でモデルを非表示にし、`モデル名=表示名` でモデルの表示名をカスタマイズし、カンマで区切ります。
 
-Azure モードでは、`modelName@azure=deploymentName` 形式でモデル名とデプロイ名（deploy-name）を設定できます。
+Azure モードでは、`modelName@Azure=deploymentName` 形式でモデル名とデプロイ名（deploy-name）を設定できます。
-> 例：`+gpt-3.5-turbo@azure=gpt35` この設定でモデルリストに `gpt35(Azure)` のオプションが表示されます。
+> 例：`+gpt-3.5-turbo@Azure=gpt35` この設定でモデルリストに `gpt35(Azure)` のオプションが表示されます。
 
 ByteDance モードでは、`modelName@bytedance=deploymentName` 形式でモデル名とデプロイ名（deploy-name）を設定できます。
 > 例: `+Doubao-lite-4k@bytedance=ep-xxxxx-xxx` この設定でモデルリストに `Doubao-lite-4k(ByteDance)` のオプションが表示されます。

app/api/openai.ts

@@ -6,7 +6,7 @@ import { NextRequest, NextResponse } from "next/server";
 import { auth } from "./auth";
 import { requestOpenai } from "./common";
 
-const ALLOWD_PATH = new Set(Object.values(OpenaiPath));
+const ALLOWED_PATH = new Set(Object.values(OpenaiPath));
 
 function getModels(remoteModelRes: OpenAIListModelResponse) {
   const config = getServerSideConfig();
@@ -34,7 +34,7 @@ export async function handle(
 
   const subpath = params.path.join("/");
 
-  if (!ALLOWD_PATH.has(subpath)) {
+  if (!ALLOWED_PATH.has(subpath)) {
     console.log("[OpenAI Route] forbidden path ", subpath);
     return NextResponse.json(
       {

app/client/api.ts

@@ -231,7 +231,7 @@ export function getHeaders(ignoreHeaders: boolean = false) {
 
   function getConfig() {
     const modelConfig = chatStore.currentSession().mask.modelConfig;
-    const isGoogle = modelConfig.providerName == ServiceProvider.Google;
+    const isGoogle = modelConfig.providerName === ServiceProvider.Google;
     const isAzure = modelConfig.providerName === ServiceProvider.Azure;
     const isAnthropic = modelConfig.providerName === ServiceProvider.Anthropic;
     const isBaidu = modelConfig.providerName == ServiceProvider.Baidu;

app/client/platforms/alibaba.ts

@@ -23,6 +23,7 @@ import {
 import { prettyObject } from "@/app/utils/format";
 import { getClientConfig } from "@/app/config/client";
 import { getMessageTextContent } from "@/app/utils";
+import { fetch } from "@/app/utils/stream";
 
 export interface OpenAIListModelResponse {
   object: string;
@@ -178,6 +179,7 @@ export class QwenApi implements LLMApi {
         controller.signal.onabort = finish;
 
         fetchEventSource(chatPath, {
+          fetch: fetch as any,
           ...chatPayload,
           async onopen(res) {
             clearTimeout(requestTimeoutId);

app/client/platforms/anthropic.ts

@@ -8,7 +8,7 @@ import {
   ChatMessageTool,
 } from "@/app/store";
 import { getClientConfig } from "@/app/config/client";
-import { DEFAULT_API_HOST } from "@/app/constant";
+import { ANTHROPIC_BASE_URL } from "@/app/constant";
 import { getMessageTextContent, isVisionModel } from "@/app/utils";
 import { preProcessImageContent, stream } from "@/app/utils/chat";
 import { cloudflareAIGatewayUrl } from "@/app/utils/cloudflare";
@@ -388,9 +388,7 @@ export class ClaudeApi implements LLMApi {
     if (baseUrl.trim().length === 0) {
       const isApp = !!getClientConfig()?.isApp;
 
-      baseUrl = isApp
+      baseUrl = isApp ? ANTHROPIC_BASE_URL : ApiPath.Anthropic;
-        ? DEFAULT_API_HOST + "/api/proxy/anthropic"
-        : ApiPath.Anthropic;
     }
 
     if (!baseUrl.startsWith("http") && !baseUrl.startsWith("/api")) {

app/client/platforms/baidu.ts

@@ -24,6 +24,7 @@ import {
 import { prettyObject } from "@/app/utils/format";
 import { getClientConfig } from "@/app/config/client";
 import { getMessageTextContent } from "@/app/utils";
+import { fetch } from "@/app/utils/stream";
 
 export interface OpenAIListModelResponse {
   object: string;
@@ -197,6 +198,7 @@ export class ErnieApi implements LLMApi {
         controller.signal.onabort = finish;
 
         fetchEventSource(chatPath, {
+          fetch: fetch as any,
           ...chatPayload,
           async onopen(res) {
             clearTimeout(requestTimeoutId);

app/client/platforms/bytedance.ts

@@ -23,6 +23,7 @@ import {
 import { prettyObject } from "@/app/utils/format";
 import { getClientConfig } from "@/app/config/client";
 import { getMessageTextContent } from "@/app/utils";
+import { fetch } from "@/app/utils/stream";
 
 export interface OpenAIListModelResponse {
   object: string;
@@ -165,6 +166,7 @@ export class DoubaoApi implements LLMApi {
         controller.signal.onabort = finish;
 
         fetchEventSource(chatPath, {
+          fetch: fetch as any,
           ...chatPayload,
           async onopen(res) {
             clearTimeout(requestTimeoutId);

app/client/platforms/google.ts

@@ -16,7 +16,7 @@ import {
 } from "@/app/store";
 import { stream } from "@/app/utils/chat";
 import { getClientConfig } from "@/app/config/client";
-import { DEFAULT_API_HOST } from "@/app/constant";
+import { GEMINI_BASE_URL } from "@/app/constant";
 
 import {
   getMessageTextContent,
@@ -26,6 +26,7 @@ import {
 import { preProcessImageContent } from "@/app/utils/chat";
 import { nanoid } from "nanoid";
 import { RequestPayload } from "./openai";
+import { fetch } from "@/app/utils/stream";
 
 export class GeminiProApi implements LLMApi {
   path(path: string): string {
@@ -38,7 +39,7 @@ export class GeminiProApi implements LLMApi {
 
     const isApp = !!getClientConfig()?.isApp;
     if (baseUrl.length === 0) {
-      baseUrl = isApp ? DEFAULT_API_HOST + `/api/proxy/google` : ApiPath.Google;
+      baseUrl = isApp ? GEMINI_BASE_URL : ApiPath.Google;
     }
     if (baseUrl.endsWith("/")) {
       baseUrl = baseUrl.slice(0, baseUrl.length - 1);

app/client/platforms/iflytek.ts

@@ -1,7 +1,7 @@
 "use client";
 import {
   ApiPath,
-  DEFAULT_API_HOST,
+  IFLYTEK_BASE_URL,
   Iflytek,
   REQUEST_TIMEOUT_MS,
 } from "@/app/constant";
@@ -22,6 +22,7 @@ import {
 import { prettyObject } from "@/app/utils/format";
 import { getClientConfig } from "@/app/config/client";
 import { getMessageTextContent } from "@/app/utils";
+import { fetch } from "@/app/utils/stream";
 
 import { RequestPayload } from "./openai";
 
@@ -40,7 +41,7 @@ export class SparkApi implements LLMApi {
     if (baseUrl.length === 0) {
       const isApp = !!getClientConfig()?.isApp;
       const apiPath = ApiPath.Iflytek;
-      baseUrl = isApp ? DEFAULT_API_HOST + "/proxy" + apiPath : apiPath;
+      baseUrl = isApp ? IFLYTEK_BASE_URL : apiPath;
     }
 
     if (baseUrl.endsWith("/")) {
@@ -149,6 +150,7 @@ export class SparkApi implements LLMApi {
         controller.signal.onabort = finish;
 
         fetchEventSource(chatPath, {
+          fetch: fetch as any,
           ...chatPayload,
           async onopen(res) {
             clearTimeout(requestTimeoutId);

app/client/platforms/moonshot.ts

@@ -2,7 +2,7 @@
 // azure and openai, using same models. so using same LLMApi.
 import {
   ApiPath,
-  DEFAULT_API_HOST,
+  MOONSHOT_BASE_URL,
   Moonshot,
   REQUEST_TIMEOUT_MS,
 } from "@/app/constant";
@@ -40,7 +40,7 @@ export class MoonshotApi implements LLMApi {
     if (baseUrl.length === 0) {
       const isApp = !!getClientConfig()?.isApp;
       const apiPath = ApiPath.Moonshot;
-      baseUrl = isApp ? DEFAULT_API_HOST + "/proxy" + apiPath : apiPath;
+      baseUrl = isApp ? MOONSHOT_BASE_URL : apiPath;
     }
 
     if (baseUrl.endsWith("/")) {

app/client/platforms/openai.ts

@@ -2,7 +2,7 @@
 // azure and openai, using same models. so using same LLMApi.
 import {
   ApiPath,
-  DEFAULT_API_HOST,
+  OPENAI_BASE_URL,
   DEFAULT_MODELS,
   OpenaiPath,
   Azure,
@@ -98,7 +98,7 @@ export class ChatGPTApi implements LLMApi {
     if (baseUrl.length === 0) {
       const isApp = !!getClientConfig()?.isApp;
       const apiPath = isAzure ? ApiPath.Azure : ApiPath.OpenAI;
-      baseUrl = isApp ? DEFAULT_API_HOST + "/proxy" + apiPath : apiPath;
+      baseUrl = isApp ? OPENAI_BASE_URL : apiPath;
     }
 
     if (baseUrl.endsWith("/")) {

app/client/platforms/tencent.ts

@@ -1,5 +1,5 @@
 "use client";
-import { ApiPath, DEFAULT_API_HOST, REQUEST_TIMEOUT_MS } from "@/app/constant";
+import { ApiPath, TENCENT_BASE_URL, REQUEST_TIMEOUT_MS } from "@/app/constant";
 import { useAccessStore, useAppConfig, useChatStore } from "@/app/store";
 
 import {
@@ -22,6 +22,7 @@ import mapKeys from "lodash-es/mapKeys";
 import mapValues from "lodash-es/mapValues";
 import isArray from "lodash-es/isArray";
 import isObject from "lodash-es/isObject";
+import { fetch } from "@/app/utils/stream";
 
 export interface OpenAIListModelResponse {
   object: string;
@@ -70,9 +71,7 @@ export class HunyuanApi implements LLMApi {
 
     if (baseUrl.length === 0) {
       const isApp = !!getClientConfig()?.isApp;
-      baseUrl = isApp
+      baseUrl = isApp ? TENCENT_BASE_URL : ApiPath.Tencent;
-        ? DEFAULT_API_HOST + "/api/proxy/tencent"
-        : ApiPath.Tencent;
     }
 
     if (baseUrl.endsWith("/")) {
@@ -179,6 +178,7 @@ export class HunyuanApi implements LLMApi {
         controller.signal.onabort = finish;
 
         fetchEventSource(chatPath, {
+          fetch: fetch as any,
           ...chatPayload,
           async onopen(res) {
             clearTimeout(requestTimeoutId);

app/components/chat.tsx

@@ -1815,6 +1815,7 @@ function _Chat() {
                       {message?.tools?.map((tool) => (
                         <div
                           key={tool.id}
+                          title={tool?.errorMsg}
                           className={styles["chat-message-tool"]}
                         >
                           {tool.isError === false ? (

app/components/markdown.tsx

@@ -207,23 +207,6 @@ function CustomCode(props: { children: any; className?: string }) {
   );
 }
 
-function escapeDollarNumber(text: string) {
-  let escapedText = "";
-
-  for (let i = 0; i < text.length; i += 1) {
-    let char = text[i];
-    const nextChar = text[i + 1] || " ";
-
-    if (char === "$" && nextChar >= "0" && nextChar <= "9") {
-      char = "\\$";
-    }
-
-    escapedText += char;
-  }
-
-  return escapedText;
-}
-
 function escapeBrackets(text: string) {
   const pattern =
     /(```[\s\S]*?```|`.*?`)|\\\[([\s\S]*?[^\\])\\\]|\\\((.*?)\\\)/g;
@@ -261,7 +244,7 @@ function tryWrapHtmlCode(text: string) {
 
 function _MarkDownContent(props: { content: string }) {
   const escapedContent = useMemo(() => {
-    return tryWrapHtmlCode(escapeBrackets(escapeDollarNumber(props.content)));
+    return tryWrapHtmlCode(escapeBrackets(props.content));
   }, [props.content]);
 
   return (

app/constant.ts

@@ -11,7 +11,6 @@ export const RUNTIME_CONFIG_DOM = "danger-runtime-config";
 
 export const STABILITY_BASE_URL = "https://api.stability.ai";
 
-export const DEFAULT_API_HOST = "https://api.nextchat.dev";
 export const OPENAI_BASE_URL = "https://api.openai.com";
 export const ANTHROPIC_BASE_URL = "https://api.anthropic.com";
 

app/store/access.ts

@@ -1,9 +1,18 @@
 import {
-  ApiPath,
-  DEFAULT_API_HOST,
   GoogleSafetySettingsThreshold,
   ServiceProvider,
   StoreKey,
+  ApiPath,
+  OPENAI_BASE_URL,
+  ANTHROPIC_BASE_URL,
+  GEMINI_BASE_URL,
+  BAIDU_BASE_URL,
+  BYTEDANCE_BASE_URL,
+  ALIBABA_BASE_URL,
+  TENCENT_BASE_URL,
+  MOONSHOT_BASE_URL,
+  STABILITY_BASE_URL,
+  IFLYTEK_BASE_URL,
 } from "../constant";
 import { getHeaders } from "../client/api";
 import { getClientConfig } from "../config/client";
@@ -15,45 +24,25 @@ let fetchState = 0; // 0 not fetch, 1 fetching, 2 done
 
 const isApp = getClientConfig()?.buildMode === "export";
 
-const DEFAULT_OPENAI_URL = isApp
+const DEFAULT_OPENAI_URL = isApp ? OPENAI_BASE_URL : ApiPath.OpenAI;
-  ? DEFAULT_API_HOST + "/api/proxy/openai"
-  : ApiPath.OpenAI;
 
-const DEFAULT_GOOGLE_URL = isApp
+const DEFAULT_GOOGLE_URL = isApp ? GEMINI_BASE_URL : ApiPath.Google;
-  ? DEFAULT_API_HOST + "/api/proxy/google"
-  : ApiPath.Google;
 
-const DEFAULT_ANTHROPIC_URL = isApp
+const DEFAULT_ANTHROPIC_URL = isApp ? ANTHROPIC_BASE_URL : ApiPath.Anthropic;
-  ? DEFAULT_API_HOST + "/api/proxy/anthropic"
-  : ApiPath.Anthropic;
 
-const DEFAULT_BAIDU_URL = isApp
+const DEFAULT_BAIDU_URL = isApp ? BAIDU_BASE_URL : ApiPath.Baidu;
-  ? DEFAULT_API_HOST + "/api/proxy/baidu"
-  : ApiPath.Baidu;
 
-const DEFAULT_BYTEDANCE_URL = isApp
+const DEFAULT_BYTEDANCE_URL = isApp ? BYTEDANCE_BASE_URL : ApiPath.ByteDance;
-  ? DEFAULT_API_HOST + "/api/proxy/bytedance"
-  : ApiPath.ByteDance;
 
-const DEFAULT_ALIBABA_URL = isApp
+const DEFAULT_ALIBABA_URL = isApp ? ALIBABA_BASE_URL : ApiPath.Alibaba;
-  ? DEFAULT_API_HOST + "/api/proxy/alibaba"
-  : ApiPath.Alibaba;
 
-const DEFAULT_TENCENT_URL = isApp
+const DEFAULT_TENCENT_URL = isApp ? TENCENT_BASE_URL : ApiPath.Tencent;
-  ? DEFAULT_API_HOST + "/api/proxy/tencent"
-  : ApiPath.Tencent;
 
-const DEFAULT_MOONSHOT_URL = isApp
+const DEFAULT_MOONSHOT_URL = isApp ? MOONSHOT_BASE_URL : ApiPath.Moonshot;
-  ? DEFAULT_API_HOST + "/api/proxy/moonshot"
-  : ApiPath.Moonshot;
 
-const DEFAULT_STABILITY_URL = isApp
+const DEFAULT_STABILITY_URL = isApp ? STABILITY_BASE_URL : ApiPath.Stability;
-  ? DEFAULT_API_HOST + "/api/proxy/stability"
-  : ApiPath.Stability;
 
-const DEFAULT_IFLYTEK_URL = isApp
+const DEFAULT_IFLYTEK_URL = isApp ? IFLYTEK_BASE_URL : ApiPath.Iflytek;
-  ? DEFAULT_API_HOST + "/api/proxy/iflytek"
-  : ApiPath.Iflytek;
 
 const DEFAULT_ACCESS_STATE = {
   accessCode: "",

app/store/chat.ts

@@ -16,6 +16,9 @@ import {
   DEFAULT_SYSTEM_TEMPLATE,
   KnowledgeCutOffDate,
   StoreKey,
+  SUMMARIZE_MODEL,
+  GEMINI_SUMMARIZE_MODEL,
+  ServiceProvider,
 } from "../constant";
 import Locale, { getLang } from "../locales";
 import { isDalle3, safeLocalStorage } from "../utils";
@@ -23,6 +26,8 @@ import { prettyObject } from "../utils/format";
 import { createPersistStore } from "../utils/store";
 import { estimateTokenLength } from "../utils/token";
 import { ModelConfig, ModelType, useAppConfig } from "./config";
+import { useAccessStore } from "./access";
+import { collectModelsWithDefaultModel } from "../utils/model";
 import { createEmptyMask, Mask } from "./mask";
 
 const localStorage = safeLocalStorage();
@@ -37,6 +42,7 @@ export type ChatMessageTool = {
   };
   content?: string;
   isError?: boolean;
+  errorMsg?: string;
 };
 
 export type ChatMessage = RequestMessage & {
@@ -102,6 +108,35 @@ function createEmptySession(): ChatSession {
   };
 }
 
+function getSummarizeModel(
+  currentModel: string,
+  providerName: string,
+): string[] {
+  // if it is using gpt-* models, force to use 4o-mini to summarize
+  if (currentModel.startsWith("gpt") || currentModel.startsWith("chatgpt")) {
+    const configStore = useAppConfig.getState();
+    const accessStore = useAccessStore.getState();
+    const allModel = collectModelsWithDefaultModel(
+      configStore.models,
+      [configStore.customModels, accessStore.customModels].join(","),
+      accessStore.defaultModel,
+    );
+    const summarizeModel = allModel.find(
+      (m) => m.name === SUMMARIZE_MODEL && m.available,
+    );
+    if (summarizeModel) {
+      return [
+        summarizeModel.name,
+        summarizeModel.provider?.providerName as string,
+      ];
+    }
+  }
+  if (currentModel.startsWith("gemini")) {
+    return [GEMINI_SUMMARIZE_MODEL, ServiceProvider.Google];
+  }
+  return [currentModel, providerName];
+}
+
 function countMessages(msgs: ChatMessage[]) {
   return msgs.reduce(
     (pre, cur) => pre + estimateTokenLength(getMessageTextContent(cur)),
@@ -578,8 +613,14 @@ export const useChatStore = createPersistStore(
           return;
         }
 
-        const providerName = modelConfig.compressProviderName;
+        // if not config compressModel, then using getSummarizeModel
-        const api: ClientApi = getClientApi(providerName);
+        const [model, providerName] = modelConfig.compressModel
+          ? [modelConfig.compressModel, modelConfig.compressProviderName]
+          : getSummarizeModel(
+              session.mask.modelConfig.model,
+              session.mask.modelConfig.providerName,
+            );
+        const api: ClientApi = getClientApi(providerName as ServiceProvider);
 
         // remove error messages if any
         const messages = session.messages;
@@ -610,7 +651,7 @@ export const useChatStore = createPersistStore(
           api.llm.chat({
             messages: topicMessages,
             config: {
-              model: modelConfig.compressModel,
+              model,
               stream: false,
               providerName,
             },
@@ -674,7 +715,8 @@ export const useChatStore = createPersistStore(
             config: {
               ...modelcfg,
               stream: true,
-              model: modelConfig.compressModel,
+              model,
+              providerName,
             },
             onUpdate(message) {
               session.memoryPrompt = message;
@@ -727,7 +769,7 @@ export const useChatStore = createPersistStore(
   },
   {
     name: StoreKey.Chat,
-    version: 3.2,
+    version: 3.3,
     migrate(persistedState, version) {
       const state = persistedState as any;
       const newState = JSON.parse(
@@ -783,6 +825,14 @@ export const useChatStore = createPersistStore(
             config.modelConfig.compressProviderName;
         });
       }
+      // revert default summarize model for every session
+      if (version < 3.3) {
+        newState.sessions.forEach((s) => {
+          const config = useAppConfig.getState();
+          s.mask.modelConfig.compressModel = "";
+          s.mask.modelConfig.compressProviderName = "";
+        });
+      }
 
       return newState as any;
     },

app/store/config.ts

@@ -71,8 +71,8 @@ export const DEFAULT_CONFIG = {
     sendMemory: true,
     historyMessageCount: 4,
     compressMessageLengthThreshold: 1000,
-    compressModel: "gpt-4o-mini" as ModelType,
+    compressModel: "",
-    compressProviderName: "OpenAI" as ServiceProvider,
+    compressProviderName: "",
     enableInjectSystemPrompts: true,
     template: config?.template ?? DEFAULT_INPUT_TEMPLATE,
     size: "1024x1024" as DalleSize,
@@ -178,7 +178,7 @@ export const useAppConfig = createPersistStore(
   }),
   {
     name: StoreKey.Config,
-    version: 4,
+    version: 4.1,
 
     merge(persistedState, currentState) {
       const state = persistedState as ChatConfig | undefined;
@@ -231,7 +231,7 @@ export const useAppConfig = createPersistStore(
             : config?.template ?? DEFAULT_INPUT_TEMPLATE;
       }
 
-      if (version < 4) {
+      if (version < 4.1) {
         state.modelConfig.compressModel =
           DEFAULT_CONFIG.modelConfig.compressModel;
         state.modelConfig.compressProviderName =

app/store/plugin.ts

@@ -7,7 +7,7 @@ import yaml from "js-yaml";
 import { adapter, getOperationId } from "../utils";
 import { useAccessStore } from "./access";
 
-const isApp = getClientConfig()?.isApp;
+const isApp = getClientConfig()?.isApp !== false;
 
 export type Plugin = {
   id: string;

app/store/sync.ts

@@ -12,7 +12,6 @@ import { downloadAs, readFromFile } from "../utils";
 import { showToast } from "../components/ui-lib";
 import Locale from "../locales";
 import { createSyncClient, ProviderType } from "../utils/cloud";
-import { corsPath } from "../utils/cors";
 
 export interface WebDavConfig {
   server: string;
@@ -26,7 +25,7 @@ export type SyncStore = GetStoreState<typeof useSyncStore>;
 const DEFAULT_SYNC_STATE = {
   provider: ProviderType.WebDAV,
   useProxy: true,
-  proxyUrl: corsPath(ApiPath.Cors),
+  proxyUrl: ApiPath.Cors as string,
 
   webdav: {
     endpoint: "",

app/utils.ts

@@ -2,8 +2,9 @@ import { useEffect, useState } from "react";
 import { showToast } from "./components/ui-lib";
 import Locale from "./locales";
 import { RequestMessage } from "./client/api";
-import { ServiceProvider, REQUEST_TIMEOUT_MS } from "./constant";
+import { ServiceProvider } from "./constant";
-import { fetch as tauriFetch, ResponseType } from "@tauri-apps/api/http";
+// import { fetch as tauriFetch, ResponseType } from "@tauri-apps/api/http";
+import { fetch as tauriStreamFetch } from "./utils/stream";
 
 export function trimTopic(topic: string) {
   // Fix an issue where double quotes still show in the Indonesian language
@@ -295,30 +296,23 @@ export function fetch(
   options?: Record<string, unknown>,
 ): Promise<any> {
   if (window.__TAURI__) {
-    const payload = options?.body || options?.data;
+    return tauriStreamFetch(url, options);
-    return tauriFetch(url, {
-      ...options,
-      body:
-        payload &&
-        ({
-          type: "Text",
-          payload,
-        } as any),
-      timeout: ((options?.timeout as number) || REQUEST_TIMEOUT_MS) / 1000,
-      responseType:
-        options?.responseType == "text" ? ResponseType.Text : ResponseType.JSON,
-    } as any);
   }
   return window.fetch(url, options);
 }
 
 export function adapter(config: Record<string, unknown>) {
-  const { baseURL, url, params, ...rest } = config;
+  const { baseURL, url, params, data: body, ...rest } = config;
   const path = baseURL ? `${baseURL}${url}` : url;
   const fetchUrl = params
     ? `${path}?${new URLSearchParams(params as any).toString()}`
     : path;
-  return fetch(fetchUrl as string, { ...rest, responseType: "text" });
+  return fetch(fetchUrl as string, { ...rest, body }).then((res) => {
+    const { status, headers, statusText } = res;
+    return res
+      .text()
+      .then((data: string) => ({ status, statusText, headers, data }));
+  });
 }
 
 export function safeLocalStorage(): {

app/utils/chat.ts

@@ -10,6 +10,7 @@ import {
   fetchEventSource,
 } from "@fortaine/fetch-event-source";
 import { prettyObject } from "./format";
+import { fetch as tauriFetch } from "./stream";
 
 export function compressImage(file: Blob, maxSize: number): Promise<string> {
   return new Promise((resolve, reject) => {
@@ -221,7 +222,7 @@ export function stream(
               ),
             )
               .then((res) => {
-                const content = JSON.stringify(res.data);
+                let content = res.data || res?.statusText;
                 if (res.status >= 300) {
                   return Promise.reject(content);
                 }
@@ -236,7 +237,11 @@ export function stream(
                 return content;
               })
               .catch((e) => {
-                options?.onAfterTool?.({ ...tool, isError: true });
+                options?.onAfterTool?.({
+                  ...tool,
+                  isError: true,
+                  errorMsg: e.toString(),
+                });
                 return e.toString();
               })
               .then((content) => ({
@@ -288,6 +293,7 @@ export function stream(
       REQUEST_TIMEOUT_MS,
     );
     fetchEventSource(chatPath, {
+      fetch: tauriFetch as any,
       ...chatPayload,
       async onopen(res) {
         clearTimeout(requestTimeoutId);

app/utils/cors.ts

@@ -1,19 +0,0 @@
-import { getClientConfig } from "../config/client";
-import { DEFAULT_API_HOST } from "../constant";
-
-export function corsPath(path: string) {
-  const baseUrl = getClientConfig()?.isApp ? `${DEFAULT_API_HOST}` : "";
-
-  if (baseUrl === "" && path === "") {
-    return "";
-  }
-  if (!path.startsWith("/")) {
-    path = "/" + path;
-  }
-
-  if (!path.endsWith("/")) {
-    path += "/";
-  }
-
-  return `${baseUrl}${path}`;
-}

app/utils/stream.ts

@@ -0,0 +1,107 @@
+// using tauri command to send request
+// see src-tauri/src/stream.rs, and src-tauri/src/main.rs
+// 1. invoke('stream_fetch', {url, method, headers, body}), get response with headers.
+// 2. listen event: `stream-response` multi times to get body
+
+type ResponseEvent = {
+  id: number;
+  payload: {
+    request_id: number;
+    status?: number;
+    chunk?: number[];
+  };
+};
+
+type StreamResponse = {
+  request_id: number;
+  status: number;
+  status_text: string;
+  headers: Record<string, string>;
+};
+
+export function fetch(url: string, options?: RequestInit): Promise<any> {
+  if (window.__TAURI__) {
+    const {
+      signal,
+      method = "GET",
+      headers: _headers = {},
+      body = [],
+    } = options || {};
+    let unlisten: Function | undefined;
+    let setRequestId: Function | undefined;
+    const requestIdPromise = new Promise((resolve) => (setRequestId = resolve));
+    const ts = new TransformStream();
+    const writer = ts.writable.getWriter();
+
+    let closed = false;
+    const close = () => {
+      if (closed) return;
+      closed = true;
+      unlisten && unlisten();
+      writer.ready.then(() => {
+        writer.close().catch((e) => console.error(e));
+      });
+    };
+
+    if (signal) {
+      signal.addEventListener("abort", () => close());
+    }
+    // @ts-ignore 2. listen response multi times, and write to Response.body
+    window.__TAURI__.event
+      .listen("stream-response", (e: ResponseEvent) =>
+        requestIdPromise.then((request_id) => {
+          const { request_id: rid, chunk, status } = e?.payload || {};
+          if (request_id != rid) {
+            return;
+          }
+          if (chunk) {
+            writer.ready.then(() => {
+              writer.write(new Uint8Array(chunk));
+            });
+          } else if (status === 0) {
+            // end of body
+            close();
+          }
+        }),
+      )
+      .then((u: Function) => (unlisten = u));
+
+    const headers: Record<string, string> = {
+      Accept: "application/json, text/plain, */*",
+      "Accept-Language": "en-US,en;q=0.9,zh-CN;q=0.8,zh;q=0.7",
+      "User-Agent": navigator.userAgent,
+    };
+    for (const item of new Headers(_headers || {})) {
+      headers[item[0]] = item[1];
+    }
+    return window.__TAURI__
+      .invoke("stream_fetch", {
+        method: method.toUpperCase(),
+        url,
+        headers,
+        // TODO FormData
+        body:
+          typeof body === "string"
+            ? Array.from(new TextEncoder().encode(body))
+            : [],
+      })
+      .then((res: StreamResponse) => {
+        const { request_id, status, status_text: statusText, headers } = res;
+        setRequestId?.(request_id);
+        const response = new Response(ts.readable, {
+          status,
+          statusText,
+          headers,
+        });
+        if (status >= 300) {
+          setTimeout(close, 100);
+        }
+        return response;
+      })
+      .catch((e) => {
+        console.error("stream error", e);
+        throw e;
+      });
+  }
+  return window.fetch(url, options);
+}

src-tauri/Cargo.lock

@@ -348,9 +348,9 @@ checksum = "14c189c53d098945499cdfa7ecc63567cf3886b3332b312a5b4585d8d3a6a610"
 
 [[package]]
 name = "bytes"
-version = "1.4.0"
+version = "1.7.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "89b2fd2a0dcf38d7971e2194b6b6eebab45ae01067456a7fd93d5547a61b70be"
+checksum = "428d9aa8fbc0670b7b8d6030a7fadd0f86151cae55e4dbbece15f3780a3dfaf3"
 dependencies = [
  "serde",
 ]
@@ -942,9 +942,9 @@ checksum = "00b0228411908ca8685dba7fc2cdd70ec9990a6e753e89b6ac91a84c40fbaf4b"
 
 [[package]]
 name = "form_urlencoded"
-version = "1.1.0"
+version = "1.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a9c384f161156f5260c24a097c56119f9be8c798586aecc13afbcbe7b7e26bf8"
+checksum = "e13624c2627564efccf4934284bdd98cbaa14e79b0b5a141218e507b3a823456"
 dependencies = [
  "percent-encoding",
 ]
@@ -970,9 +970,9 @@ dependencies = [
 
 [[package]]
 name = "futures-core"
-version = "0.3.28"
+version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4bca583b7e26f571124fe5b7561d49cb2868d79116cfa0eefce955557c6fee8c"
+checksum = "dfc6580bb841c5a68e9ef15c77ccc837b40a7504914d52e47b8b0e9bbda25a1d"
 
 [[package]]
 name = "futures-executor"
@@ -987,9 +987,9 @@ dependencies = [
 
 [[package]]
 name = "futures-io"
-version = "0.3.28"
+version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4fff74096e71ed47f8e023204cfd0aa1289cd54ae5430a9523be060cdb849964"
+checksum = "a44623e20b9681a318efdd71c299b6b222ed6f231972bfe2f224ebad6311f0c1"
 
 [[package]]
 name = "futures-lite"
@@ -1008,9 +1008,9 @@ dependencies = [
 
 [[package]]
 name = "futures-macro"
-version = "0.3.28"
+version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "89ca545a94061b6365f2c7355b4b32bd20df3ff95f02da9329b34ccc3bd6ee72"
+checksum = "87750cf4b7a4c0625b1529e4c543c2182106e4dedc60a2a6455e00d212c489ac"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -1019,21 +1019,21 @@ dependencies = [
 
 [[package]]
 name = "futures-sink"
-version = "0.3.29"
+version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e36d3378ee38c2a36ad710c5d30c2911d752cb941c00c72dbabfb786a7970817"
+checksum = "9fb8e00e87438d937621c1c6269e53f536c14d3fbd6a042bb24879e57d474fb5"
 
 [[package]]
 name = "futures-task"
-version = "0.3.28"
+version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "76d3d132be6c0e6aa1534069c705a74a5997a356c0dc2f86a47765e5617c5b65"
+checksum = "38d84fa142264698cdce1a9f9172cf383a0c82de1bddcf3092901442c4097004"
 
 [[package]]
 name = "futures-util"
-version = "0.3.28"
+version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "26b01e40b772d54cf6c6d721c1d1abd0647a0106a12ecaa1c186273392a69533"
+checksum = "3d6401deb83407ab3da39eba7e33987a73c3df0c82b4bb5813ee871c19c41d48"
 dependencies = [
  "futures-core",
  "futures-io",
@@ -1555,9 +1555,9 @@ checksum = "b9e0384b61958566e926dc50660321d12159025e767c18e043daf26b70104c39"
 
 [[package]]
 name = "idna"
-version = "0.3.0"
+version = "0.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e14ddfc70884202db2244c223200c204c2bda1bc6e0998d11b5e024d657209e6"
+checksum = "634d9b1461af396cad843f47fdba5597a4f9e6ddd4bfb6ff5d85028c25cb12f6"
 dependencies = [
  "unicode-bidi",
  "unicode-normalization",
@@ -1986,6 +1986,10 @@ checksum = "e4a24736216ec316047a1fc4252e27dabb04218aa4a3f37c6e7ddbf1f9782b54"
 name = "nextchat"
 version = "0.1.0"
 dependencies = [
+ "bytes",
+ "futures-util",
+ "percent-encoding",
+ "reqwest",
  "serde",
  "serde_json",
  "tauri",
@@ -2281,9 +2285,9 @@ checksum = "8835116a5c179084a830efb3adc117ab007512b535bc1a21c991d3b32a6b44dd"
 
 [[package]]
 name = "percent-encoding"
-version = "2.2.0"
+version = "2.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "478c572c3d73181ff3c2539045f6eb99e5491218eae919370993b890cdbdd98e"
+checksum = "e3148f5046208a5d56bcfc03053e3ca6334e51da8dfb19b6cdc8b306fae3283e"
 
 [[package]]
 name = "phf"
@@ -2545,9 +2549,9 @@ checksum = "dc375e1527247fe1a97d8b7156678dfe7c1af2fc075c9a4db3690ecd2a148068"
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.58"
+version = "1.0.86"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fa1fb82fc0c281dd9671101b66b771ebbe1eaf967b96ac8740dcba4b70005ca8"
+checksum = "5e719e8df665df0d1c8fbfd238015744736151d4445ec0836b8e628aae103b77"
 dependencies = [
  "unicode-ident",
 ]
@@ -3889,9 +3893,9 @@ checksum = "1dd624098567895118886609431a7c3b8f516e41d30e0643f03d94592a147e36"
 
 [[package]]
 name = "url"
-version = "2.3.1"
+version = "2.5.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0d68c799ae75762b8c3fe375feb6600ef5602c883c5d21eb51c09f22b83c4643"
+checksum = "22784dbdf76fdde8af1aeda5622b546b422b6fc585325248a2bf9f5e41e94d6c"
 dependencies = [
  "form_urlencoded",
  "idna",

src-tauri/Cargo.toml

@@ -37,6 +37,10 @@ tauri = { version = "1.5.4", features = [ "http-all",
     "window-unminimize",
 ] }
 tauri-plugin-window-state = { git = "https://github.com/tauri-apps/plugins-workspace", branch = "v1" }
+percent-encoding = "2.3.1"
+reqwest = "0.11.18"
+futures-util = "0.3.30"
+bytes = "1.7.2"
 
 [features]
 # this feature is used for production builds or when `devPath` points to the filesystem and the built-in dev server is disabled.

src-tauri/src/main.rs

@@ -1,8 +1,11 @@
 // Prevents additional console window on Windows in release, DO NOT REMOVE!!
 #![cfg_attr(not(debug_assertions), windows_subsystem = "windows")]
 
+mod stream;
+
 fn main() {
   tauri::Builder::default()
+    .invoke_handler(tauri::generate_handler![stream::stream_fetch])
     .plugin(tauri_plugin_window_state::Builder::default().build())
     .run(tauri::generate_context!())
     .expect("error while running tauri application");

src-tauri/src/stream.rs

@@ -0,0 +1,134 @@
+//
+//
+
+use std::time::Duration;
+use std::error::Error;
+use std::sync::atomic::{AtomicU32, Ordering};
+use std::collections::HashMap;
+use futures_util::{StreamExt};
+use reqwest::Client;
+use reqwest::header::{HeaderName, HeaderMap};
+
+static REQUEST_COUNTER: AtomicU32 = AtomicU32::new(0);
+
+#[derive(Debug, Clone, serde::Serialize)]
+pub struct StreamResponse {
+  request_id: u32,
+  status: u16,
+  status_text: String,
+  headers: HashMap<String, String>
+}
+
+#[derive(Clone, serde::Serialize)]
+pub struct EndPayload {
+  request_id: u32,
+  status: u16,
+}
+
+#[derive(Clone, serde::Serialize)]
+pub struct ChunkPayload {
+  request_id: u32,
+  chunk: bytes::Bytes,
+}
+
+#[tauri::command]
+pub async fn stream_fetch(
+  window: tauri::Window,
+  method: String,
+  url: String,
+  headers: HashMap<String, String>,
+  body: Vec<u8>,
+) -> Result<StreamResponse, String> {
+
+  let event_name = "stream-response";
+  let request_id = REQUEST_COUNTER.fetch_add(1, Ordering::SeqCst);
+
+  let mut _headers = HeaderMap::new();
+  for (key, value) in &headers {
+    _headers.insert(key.parse::<HeaderName>().unwrap(), value.parse().unwrap());
+  }
+
+  // println!("method: {:?}", method);
+  // println!("url: {:?}", url);
+  // println!("headers: {:?}", headers);
+  // println!("headers: {:?}", _headers);
+
+  let method = method.parse::<reqwest::Method>().map_err(|err| format!("failed to parse method: {}", err))?;
+  let client = Client::builder()
+    .default_headers(_headers)
+    .redirect(reqwest::redirect::Policy::limited(3))
+    .connect_timeout(Duration::new(3, 0))
+    .build()
+    .map_err(|err| format!("failed to generate client: {}", err))?;
+
+  let mut request = client.request(
+    method.clone(),
+    url.parse::<reqwest::Url>().map_err(|err| format!("failed to parse url: {}", err))?
+  );
+
+  if method == reqwest::Method::POST || method == reqwest::Method::PUT || method == reqwest::Method::PATCH {
+    let body = bytes::Bytes::from(body);
+    // println!("body: {:?}", body);
+    request = request.body(body);
+  }
+
+  // println!("client: {:?}", client);
+  // println!("request: {:?}", request);
+
+  let response_future = request.send();
+
+  let res = response_future.await;
+  let response = match res {
+    Ok(res) => {
+      // get response and emit to client
+      let mut headers = HashMap::new();
+      for (name, value) in res.headers() {
+        headers.insert(
+          name.as_str().to_string(),
+          std::str::from_utf8(value.as_bytes()).unwrap().to_string()
+        );
+      }
+      let status = res.status().as_u16();
+
+      tauri::async_runtime::spawn(async move {
+        let mut stream = res.bytes_stream();
+
+        while let Some(chunk) = stream.next().await {
+          match chunk {
+            Ok(bytes) => {
+              // println!("chunk: {:?}", bytes);
+              if let Err(e) = window.emit(event_name, ChunkPayload{ request_id, chunk: bytes }) {
+                println!("Failed to emit chunk payload: {:?}", e);
+              }
+            }
+            Err(err) => {
+              println!("Error chunk: {:?}", err);
+            }
+          }
+        }
+        if let Err(e) = window.emit(event_name, EndPayload{ request_id, status: 0 }) {
+          println!("Failed to emit end payload: {:?}", e);
+        }
+      });
+
+      StreamResponse {
+        request_id,
+        status,
+        status_text: "OK".to_string(),
+        headers,
+      }
+    }
+    Err(err) => {
+      println!("Error response: {:?}", err.source().expect("REASON").to_string());
+      StreamResponse {
+        request_id,
+        status: 599,
+        status_text: err.source().expect("REASON").to_string(),
+        headers: HashMap::new(),
+      }
+    }
+  };
+  // println!("Response: {:?}", response);
+  Ok(response)
+}
+

src-tauri/tauri.conf.json

@@ -9,7 +9,7 @@
   },
   "package": {
     "productName": "NextChat",
-    "version": "2.15.3"
+    "version": "2.15.4"
   },
   "tauri": {
     "allowlist": {