yeahhe365 · octo-patch · Mar 28, 2026
diff --git a/README.md b/README.md
@@ -71,6 +71,21 @@ npm run dev
 
 ---
 
+## 🌐 Multi-Provider Support
+
+All Model Chat now supports multiple AI providers beyond Google Gemini:
+
+| Provider | Models | Context Window | Configuration |
+| :--- | :--- | :--- | :--- |
+| **Google Gemini** | Gemini 3.0, 2.5, Gemma | Up to 2M tokens | API Key or Environment Variable |
+| **MiniMax AI** | MiniMax-M2.7, MiniMax-M2.7-highspeed | 204K tokens | Settings > API Configuration |
+
+To use MiniMax models, enter your MiniMax API key in **Settings > API Configuration > MiniMax AI Configuration**, then select a MiniMax model from the model picker.
+
+> Get your MiniMax API key at [platform.minimax.chat](https://platform.minimax.chat)
+
+---
+
 ## 🛠️ 技术架构 / Technical Architecture
 
 <table width="100%">

diff --git a/all-model-chat/components/settings/SettingsContent.tsx b/all-model-chat/components/settings/SettingsContent.tsx
@@ -114,6 +114,8 @@ export const SettingsContent: React.FC<SettingsContentProps> = ({
                         setApiProxyUrl={(val) => updateSetting('apiProxyUrl', val)}
                         useApiProxy={currentSettings.useApiProxy ?? false}
                         setUseApiProxy={(val) => updateSetting('useApiProxy', val)}
+                        minimaxApiKey={currentSettings.minimaxApiKey ?? null}
+                        setMinimaxApiKey={(val) => updateSetting('minimaxApiKey', val)}
                         availableModels={availableModels}
                         t={t as any}
                     />

diff --git a/all-model-chat/components/settings/sections/ApiConfigSection.tsx b/all-model-chat/components/settings/sections/ApiConfigSection.tsx
@@ -1,6 +1,6 @@
 
 import React, { useState, useEffect } from 'react';
-import { KeyRound } from 'lucide-react';
+import { KeyRound, Sparkles } from 'lucide-react';
 import { useResponsiveValue } from '../../../hooks/useDevice';
 import { getClient } from '../../../services/api/baseApi';
 import { parseApiKeys } from '../../../utils/apiUtils';
@@ -9,6 +9,7 @@ import { ApiKeyInput } from './api-config/ApiKeyInput';
 import { ApiProxySettings } from './api-config/ApiProxySettings';
 import { ApiConnectionTester } from './api-config/ApiConnectionTester';
 import { ModelOption } from '../../../types';
+import { SETTINGS_INPUT_CLASS } from '../../../constants/appConstants';
 
 interface ApiConfigSectionProps {
   useCustomApiConfig: boolean;
@@ -19,6 +20,8 @@ interface ApiConfigSectionProps {
   setApiProxyUrl: (value: string | null) => void;
   useApiProxy: boolean;
   setUseApiProxy: (value: boolean) => void;
+  minimaxApiKey: string | null;
+  setMinimaxApiKey: (value: string | null) => void;
   availableModels: ModelOption[];
   t: (key: string) => string;
 }
@@ -42,6 +45,8 @@ export const ApiConfigSection: React.FC<ApiConfigSectionProps> = ({
   setApiProxyUrl,
   useApiProxy,
   setUseApiProxy,
+  minimaxApiKey,
+  setMinimaxApiKey,
   availableModels,
   t,
 }) => {
@@ -170,6 +175,30 @@ export const ApiConfigSection: React.FC<ApiConfigSectionProps> = ({
             </div>
         </div>
       </div>
+
+      {/* MiniMax API Configuration */}
+      <div className="mt-8 pt-6 border-t border-[var(--theme-border-primary)]">
+        <h3 className="text-base font-semibold text-[var(--theme-text-primary)] flex items-center gap-2 mb-4">
+            <Sparkles size={iconSize} className="text-[var(--theme-text-link)]" strokeWidth={1.5} />
+            {t('settingsMiniMaxConfig')}
+        </h3>
+        <p className="text-xs text-[var(--theme-text-tertiary)] mb-3">
+            {t('settingsMiniMaxHelp')}
+        </p>
+        <div className="relative">
+            <input
+                type="password"
+                value={minimaxApiKey || ''}
+                onChange={(e) => setMinimaxApiKey(e.target.value || null)}
+                placeholder={t('settingsMiniMaxKeyPlaceholder')}
+                className={`w-full px-4 py-2.5 rounded-lg text-sm border ${SETTINGS_INPUT_CLASS} focus:outline-none focus:ring-2 transition-colors duration-200`}
+                autoComplete="off"
+            />
+        </div>
+        <p className="text-xs text-[var(--theme-text-tertiary)] mt-2">
+            {t('settingsMiniMaxModelsInfo')}
+        </p>
+      </div>
     </div>
   );
 };
diff --git a/all-model-chat/constants/appConstants.ts b/all-model-chat/constants/appConstants.ts
@@ -86,6 +86,7 @@ export const DEFAULT_APP_SETTINGS: AppSettings = {
   apiKey: null,
   apiProxyUrl: "https://api-proxy.de/gemini/v1beta",
   useApiProxy: false,
+  minimaxApiKey: null,
   language: 'system',
   isStreamingEnabled: DEFAULT_IS_STREAMING_ENABLED,
   transcriptionModelId: DEFAULT_TRANSCRIPTION_MODEL_ID,

diff --git a/all-model-chat/constants/providerConstants.ts b/all-model-chat/constants/providerConstants.ts
@@ -0,0 +1,49 @@
+
+import { ModelOption } from '../types';
+
+/**
+ * Provider type identifiers.
+ * 'gemini' uses the @google/genai SDK directly.
+ * 'openai-compatible' uses the OpenAI-compatible /v1/chat/completions endpoint.
+ */
+export type ProviderType = 'gemini' | 'openai-compatible';
+
+export interface ProviderPreset {
+    name: string;
+    type: ProviderType;
+    baseUrl: string;
+    models: ModelOption[];
+    /** Model ID prefix used for detection. */
+    prefix: string;
+}
+
+export const MINIMAX_BASE_URL = 'https://api.minimax.io/v1';
+
+export const MINIMAX_MODELS: ModelOption[] = [
+    { id: 'MiniMax-M2.7', name: 'MiniMax M2.7', isPinned: true },
+    { id: 'MiniMax-M2.7-highspeed', name: 'MiniMax M2.7 Highspeed', isPinned: true },
+];
+
+/**
+ * Check if a model ID belongs to MiniMax.
+ */
+export const isMiniMaxModel = (modelId: string): boolean => {
+    return modelId.startsWith('MiniMax-');
+};
+
+/**
+ * Check if a model uses the OpenAI-compatible API path.
+ */
+export const isOpenAICompatModel = (modelId: string): boolean => {
+    return isMiniMaxModel(modelId);
+};
+
+/**
+ * Get the base URL for an OpenAI-compatible model.
+ */
+export const getOpenAICompatBaseUrl = (modelId: string): string => {
+    if (isMiniMaxModel(modelId)) {
+        return MINIMAX_BASE_URL;
+    }
+    return '';
+};
diff --git a/all-model-chat/hooks/message-sender/standard/useApiInteraction.ts b/all-model-chat/hooks/message-sender/standard/useApiInteraction.ts
@@ -2,8 +2,10 @@
 import React, { useCallback, Dispatch, SetStateAction } from 'react';
 import { AppSettings, ChatMessage, ChatSettings as IndividualChatSettings, UploadedFile } from '../../../types';
 import { createChatHistoryForApi, isGemini3Model, logService } from '../../../utils/appUtils';
+import { isOpenAICompatModel } from '../../../utils/modelHelpers';
 import { buildGenerationConfig } from '../../../services/api/baseApi';
 import { geminiServiceInstance } from '../../../services/geminiService';
+import { sendOpenAICompatMessageStream, sendOpenAICompatMessageNonStream } from '../../../services/api/openaiCompatApi';
 import { pyodideService } from '../../../services/pyodideService';
 import { isLikelyHtml } from '../../../utils/codeUtils';
 import { GetStreamHandlers } from '../types';
@@ -112,24 +114,6 @@ export const useApiInteraction = ({
         const shouldStripThinking = sessionToUpdate.hideThinkingInContext ?? appSettings.hideThinkingInContext;
         const historyForChat = await createChatHistoryForApi(baseMessagesForApi, shouldStripThinking);
 
-        const config = buildGenerationConfig(
-            activeModelId,
-            sessionToUpdate.systemInstruction,
-            { temperature: sessionToUpdate.temperature, topP: sessionToUpdate.topP },
-            sessionToUpdate.showThoughts,
-            sessionToUpdate.thinkingBudget,
-            !!sessionToUpdate.isGoogleSearchEnabled,
-            !!sessionToUpdate.isCodeExecutionEnabled,
-            !!sessionToUpdate.isUrlContextEnabled,
-            sessionToUpdate.thinkingLevel,
-            aspectRatio,
-            sessionToUpdate.isDeepSearchEnabled,
-            imageSize,
-            sessionToUpdate.safetySettings,
-            sessionToUpdate.mediaResolution,
-            !!sessionToUpdate.isLocalPythonEnabled
-        );
-
         const { streamOnError, streamOnComplete, streamOnPart, onThoughtChunk } = getStreamHandlers(
             finalSessionId,
             generationId,
@@ -149,6 +133,71 @@ export const useApiInteraction = ({
         setSessionLoading(finalSessionId, true);
         activeJobs.current.set(generationId, newAbortController);
 
+        // Route through OpenAI-compatible API for MiniMax and other non-Gemini providers
+        if (isOpenAICompatModel(activeModelId)) {
+            const minimaxKey = appSettings.minimaxApiKey;
+            if (!minimaxKey) {
+                streamOnError(new Error('MiniMax API Key is not configured. Please set it in Settings > API Configuration.'));
+                return;
+            }
+
+            const openaiConfig = {
+                temperature: sessionToUpdate.temperature,
+                topP: sessionToUpdate.topP,
+                systemInstruction: sessionToUpdate.systemInstruction,
+            };
+
+            if (appSettings.isStreamingEnabled) {
+                await sendOpenAICompatMessageStream(
+                    minimaxKey,
+                    activeModelId,
+                    historyForChat,
+                    finalParts,
+                    openaiConfig,
+                    newAbortController.signal,
+                    streamOnPart,
+                    onThoughtChunk,
+                    streamOnError,
+                    streamOnComplete,
+                    finalRole
+                );
+            } else {
+                await sendOpenAICompatMessageNonStream(
+                    minimaxKey,
+                    activeModelId,
+                    historyForChat,
+                    finalParts,
+                    openaiConfig,
+                    newAbortController.signal,
+                    streamOnError,
+                    (parts, thoughts, usage, grounding) => {
+                        for (const part of parts) streamOnPart(part);
+                        if (thoughts) onThoughtChunk(thoughts);
+                        streamOnComplete(usage, grounding);
+                    }
+                );
+            }
+            return;
+        }
+
+        const config = buildGenerationConfig(
+            activeModelId,
+            sessionToUpdate.systemInstruction,
+            { temperature: sessionToUpdate.temperature, topP: sessionToUpdate.topP },
+            sessionToUpdate.showThoughts,
+            sessionToUpdate.thinkingBudget,
+            !!sessionToUpdate.isGoogleSearchEnabled,
+            !!sessionToUpdate.isCodeExecutionEnabled,
+            !!sessionToUpdate.isUrlContextEnabled,
+            sessionToUpdate.thinkingLevel,
+            aspectRatio,
+            sessionToUpdate.isDeepSearchEnabled,
+            imageSize,
+            sessionToUpdate.safetySettings,
+            sessionToUpdate.mediaResolution,
+            !!sessionToUpdate.isLocalPythonEnabled
+        );
+
         if (appSettings.isStreamingEnabled) {
             await geminiServiceInstance.sendMessageStream(
                 keyToUse,

diff --git a/all-model-chat/hooks/message-sender/useStandardChat.ts b/all-model-chat/hooks/message-sender/useStandardChat.ts
@@ -1,6 +1,7 @@
 
 import { useCallback } from 'react';
 import { generateUniqueId, buildContentParts, getKeyForRequest, performOptimisticSessionUpdate, logService } from '../../utils/appUtils';
+import { isOpenAICompatModel } from '../../utils/modelHelpers';
 import { DEFAULT_CHAT_SETTINGS, MODELS_SUPPORTING_RAW_MODE } from '../../constants/appConstants';
 import { UploadedFile, ChatMessage } from '../../types';
 import { StandardChatProps } from './types';
@@ -66,25 +67,37 @@ export const useStandardChat = ({
             logService.info(`Fast Mode activated (One-off): Overriding thinking level to ${targetLevel}.`);
         }
 
-        const keyResult = getKeyForRequest(appSettings, settingsForApi);
-        if ('error' in keyResult) {
-            logService.error("Send message failed: API Key not configured.");
-             const errorMsg: ChatMessage = { id: generateUniqueId(), role: 'error', content: keyResult.error, timestamp: new Date() };
-             const newSessionId = generateUniqueId();
-
-             updateAndPersistSessions(prev => performOptimisticSessionUpdate(prev, {
-                 activeSessionId: null,
-                 newSessionId,
-                 newMessages: [errorMsg],
-                 settings: { ...DEFAULT_CHAT_SETTINGS, ...appSettings },
-                 appSettings,
-                 title: "API Key Error"
-             }));
-             setActiveSessionId(newSessionId);
-            return;
+        // For OpenAI-compatible providers (MiniMax), bypass Gemini key check
+        let keyToUse: string;
+        let isNewKey = false;
+        let shouldLockKey = false;
+
+        if (isOpenAICompatModel(activeModelId)) {
+            // MiniMax models use their own API key, managed in useApiInteraction
+            keyToUse = 'openai-compat-placeholder';
+            isNewKey = false;
+        } else {
+            const keyResult = getKeyForRequest(appSettings, settingsForApi);
+            if ('error' in keyResult) {
+                logService.error("Send message failed: API Key not configured.");
+                 const errorMsg: ChatMessage = { id: generateUniqueId(), role: 'error', content: keyResult.error, timestamp: new Date() };
+                 const newSessionId = generateUniqueId();
+
+                 updateAndPersistSessions(prev => performOptimisticSessionUpdate(prev, {
+                     activeSessionId: null,
+                     newSessionId,
+                     newMessages: [errorMsg],
+                     settings: { ...DEFAULT_CHAT_SETTINGS, ...appSettings },
+                     appSettings,
+                     title: "API Key Error"
+                 }));
+                 setActiveSessionId(newSessionId);
+                return;
+            }
+            keyToUse = keyResult.key;
+            isNewKey = keyResult.isNewKey;
+            shouldLockKey = isNewKey && filesToUse.some(f => f.fileUri && f.uploadState === 'active');
         }
-        const { key: keyToUse, isNewKey } = keyResult;
-        const shouldLockKey = isNewKey && filesToUse.some(f => f.fileUri && f.uploadState === 'active');
 
         const newAbortController = new AbortController();
 

diff --git a/all-model-chat/package.json b/all-model-chat/package.json
@@ -6,7 +6,9 @@
   "scripts": {
     "dev": "vite",
     "build": "vite build",
-    "preview": "vite preview"
+    "preview": "vite preview",
+    "test": "vitest run",
+    "test:watch": "vitest"
   },
   "dependencies": {
     "@formkit/auto-animate": "^0.8.2",
@@ -39,6 +41,8 @@
     "xlsx": "^0.18.5"
   },
   "devDependencies": {
+    "@testing-library/jest-dom": "^6.9.1",
+    "@testing-library/react": "^16.3.2",
     "@types/dompurify": "^3.0.5",
     "@types/katex": "^0.16.7",
     "@types/node": "^20.14.10",
@@ -47,10 +51,12 @@
     "@types/turndown": "^5.0.5",
     "@vitejs/plugin-react": "^4.3.1",
     "autoprefixer": "^10.4.19",
+    "jsdom": "^29.0.1",
     "postcss": "^8.4.39",
     "tailwindcss": "^3.4.4",
     "typescript": "^5.5.3",
     "vite": "^5.3.3",
-    "vite-plugin-static-copy": "^1.0.0"
+    "vite-plugin-static-copy": "^1.0.0",
+    "vitest": "^4.1.2"
   }
 }