docs: update README - remove Claude Code ref, add DeepSeek, remove Infini, use English commands

Author: jqknono

CHANGELOG.md

@@ -2,6 +2,16 @@
 
 All notable changes to this project will be documented in this file.
 
+## [0.7.18] - 2026-05-06
+- docs: add dashboard preview screenshot
+- feat(provider): 支持在模型返回空响应时自动切换至 Responses API
+- feat(pages): 抓取失败的 provider 不再显示套餐卡片
+- chore(metrics): 更新 OpenRouter 提供商指标数据
+- chore(assets): 更新 OpenRouter 提供商指标与定价数据
+- chore: 统一 API 密钥变量名并优化工作流推送逻辑
+- feat(provider): 更新服务商指标与定价并增强抓取校验
+- feat(provider): 更新模型服务商运行指标与定价数据
+
 ## [0.7.17] - 2026-05-06
 - feat(pages): 抓取失败的 provider 不再显示套餐卡片
 - chore(metrics): 更新 OpenRouter 提供商指标数据
@@ -42,7 +52,4 @@ All notable changes to this project will be documented in this file.
 - fix(config): 模型写回时不再默认落 apiStyle/maxInputTokens/maxOutputTokens，仅保留用户显式配置的字段
 
 ## [0.6.26] - 2026-03-31
-- feat(config): 更新供应商配置并优化模型管理
-
-## [0.6.25] - 2026-03-30
-- feat(套餐使用): 新增套餐使用状态栏与轮询功能
+- feat(config): 更新供应商配置并优化模型管理

README.md

@@ -2,14 +2,14 @@
 
 **一键切换多厂商 AI 模型，打破 Copilot 套餐限制。**
 
-支持智谱、Kimi、讯飞、火山云、Minimax、百度千帆、腾讯云、京东云、快手 KAT、X-AIO、Compshare、阿里云、Infini、七牛等国产大厂，以及**任何**兼容 OpenAI Chat、OpenAI Responses 或 Anthropic 接口风格的供应商。无需改变使用习惯，直接在 VS Code Copilot Chat 中无缝调用。
+支持智谱、Kimi、讯飞、火山引擎、MiniMax、百度千帆、腾讯云、京东云、快手 KAT、X-AIO、Compshare、阿里云、七牛、DeepSeek 等国产大厂，以及**任何**兼容 OpenAI Chat、OpenAI Responses 或 Anthropic 接口风格的供应商。无需改变使用习惯，直接在 VS Code Copilot Chat 中无缝调用。
 
 ---
 
 ## 核心特性
 
 - **多协议统一接入**：支持 OpenAI Chat（`/chat/completions`）、OpenAI Responses（`/responses`）、Anthropic（`/messages`）三种协议风格，适配任意兼容供应商。
-- **Claude Code 优先端点**：内置供应商默认使用 Anthropic 兼容端点，同时兼容 Claude Code 与 Copilot Chat。
+- **Anthropic 兼容优先**：内置供应商默认使用 Anthropic 兼容端点（`/messages`），无缝接入各类模型。
 - **零学习成本**：完全集成到 VS Code Copilot Chat，不改变任何操作习惯。
 - **灵活模型管理**：支持动态拉取 `/models` 端点，也可自定义模型列表。
 - **智能 Commit 生成**：基于 Git 变更自动生成符合 Conventional Commits 规范的提交消息。
@@ -48,10 +48,10 @@ code --install-extension techfetch-dev.coding-plans-for-copilot
 
 ### 配置
 
-1. 按 `Ctrl+Shift+P`，输入 `编码套餐: 管理编码套餐配置`
-2. 选择「选择供应商」，选择你已注册的平台（如智谱、Kimi、火山引擎等）
-3. 选择「设置 API Key」，粘贴你的 API Key
-4. 打开 Copilot Chat（`Ctrl+L`），切换到「编码套餐」提供商
+1. 按 `Ctrl+Shift+P`，输入 `Coding Plans: Manage Vendor Configuration`
+2. 选择「Select Vendor」，选择你已注册的平台（如智谱、Kimi、火山引擎等）
+3. 选择「Set API Key」，粘贴你的 API Key
+4. 打开 Copilot Chat（`Ctrl+L`），切换到「Coding Plans」提供商
 
 也可以直接编辑 `settings.json`，插件会打开设置页并定位到 `coding-plans.vendors`。
 
@@ -71,8 +71,8 @@ code --install-extension techfetch-dev.coding-plans-for-copilot
 | Kimi Overseas | `https://api.moonshot.ai/anthropic` | `https://api.moonshot.ai/v1` |
 | 阿里云（Aliyun） | `https://coding.dashscope.aliyuncs.com/apps/anthropic` | `https://coding.dashscope.aliyuncs.com/v1` |
 | 腾讯云 | `https://api.lkeap.cloud.tencent.com/coding/anthropic` | — |
-| Infini（无问芯穹） | `https://cloud.infini-ai.com` | — |
 | 七牛（Qiniu） | `https://api.qnaigc.com` | — |
+| DeepSeek | `https://api.deepseek.com/anthropic` | `https://api.deepseek.com/v1` |
 | OpenRouter | `https://openrouter.ai/api` | `https://openrouter.ai/api/v1` |
 
 如需切换到 OpenAI 兼容端点，修改供应商的 `baseUrl` 和 `defaultApiStyle` 即可。
@@ -193,7 +193,7 @@ code --install-extension techfetch-dev.coding-plans-for-copilot
 
 ### 智能 Commit 消息生成
 
-1. 按 `Ctrl+Shift+P`，输入 `编码套餐: 生成 Commit 消息`
+1. 按 `Ctrl+Shift+P`，输入 `Coding Plans: Generate Commit Message`
 2. 插件会分析当前 Git 变更，自动生成符合规范的提交消息
 3. 可选择使用的模型（默认使用当前配置的供应商）
 

README_en.md

@@ -2,14 +2,14 @@
 
 **Switch between multiple AI model vendors with one click, breaking Copilot plan limitations.**
 
-Supports domestic major vendors like Zhipu, Kimi, iFlytek, Volcano Cloud, Minimax, Baidu Qianfan, Tencent Cloud, JD Cloud, Kuaishou KAT, X-AIO, Compshare, Alibaba Cloud, Infini, Qiniu, as well as **any** vendor compatible with OpenAI Chat, OpenAI Responses, or Anthropic API styles. No need to change usage habits; seamlessly call directly in VS Code Copilot Chat.
+Supports domestic major vendors like Zhipu, Kimi, iFlytek, Volcengine, MiniMax, Baidu Qianfan, Tencent Cloud, JD Cloud, Kuaishou KAT, X-AIO, Compshare, Alibaba Cloud, Qiniu, DeepSeek, as well as **any** vendor compatible with OpenAI Chat, OpenAI Responses, or Anthropic API styles. No need to change usage habits; seamlessly call directly in VS Code Copilot Chat.
 
 ---
 
 ## Core Features
 
 - **Multi-Protocol Unified Access**: Supports OpenAI Chat (`/chat/completions`), OpenAI Responses (`/responses`), and Anthropic (`/messages`) three protocol styles, adapting to any compatible vendor.
-- **Claude Code Priority Endpoint**: Built-in vendors default to Anthropic-compatible endpoints, compatible with both Claude Code and Copilot Chat.
+- **Anthropic-Compatible Priority**: Built-in vendors default to Anthropic-compatible endpoints (`/messages`), seamlessly connecting to various models.
 - **Zero Learning Curve**: Fully integrated into VS Code Copilot Chat without changing any operational habits.
 - **Flexible Model Management**: Supports dynamic fetching from `/models` endpoint, or custom model lists.
 - **Intelligent Commit Generation**: Automatically generates Conventional Commits-compliant commit messages based on Git changes.
@@ -28,9 +28,9 @@ Supports domestic major vendors like Zhipu, Kimi, iFlytek, Volcano Cloud, Minima
 
 1. Open VS Code
 2. Press `Ctrl+Shift+X` to open the Extensions panel
-3. Type `Coding Plans for Copilot` or `编码套餐` in the search box
+3. Type `Coding Plans for Copilot` in the search box
 4. Click **Install** to install
-5. After installation, press `Ctrl+Shift+P` and type `编码套餐` to see related commands
+5. After installation, press `Ctrl+Shift+P` and type `Coding Plans` to see related commands
 
 #### Method 2: Command Line Installation
 
@@ -48,8 +48,8 @@ Click the **Install** button on the marketplace page, which will automatically o
 
 ### Configuration
 
-1. Press `Ctrl+Shift+P`, type `编码套餐: 管理编码套餐配置`
-2. Select "Select Vendor", choose the platform you've registered with (e.g., Zhipu, Kimi, Volcano Engine, etc.)
+1. Press `Ctrl+Shift+P`, type `Coding Plans: Manage Vendor Configuration`
+2. Select "Select Vendor", choose the platform you've registered with (e.g., Zhipu, Kimi, Volcengine, etc.)
 3. Select "Set API Key", paste your API Key
 4. Open Copilot Chat (`Ctrl+L`), switch to "Coding Plans" provider
 
@@ -71,8 +71,8 @@ The following vendors come with built-in default configurations and are ready to
 | Kimi Overseas | `https://api.moonshot.ai/anthropic` | `https://api.moonshot.ai/v1` |
 | Alibaba Cloud (Aliyun) | `https://coding.dashscope.aliyuncs.com/apps/anthropic` | `https://coding.dashscope.aliyuncs.com/v1` |
 | Tencent Cloud | `https://api.lkeap.cloud.tencent.com/coding/anthropic` | — |
-| Infini (无问芯穹) | `https://cloud.infini-ai.com` | — |
 | Qiniu (七牛) | `https://api.qnaigc.com` | — |
+| DeepSeek | `https://api.deepseek.com/anthropic` | `https://api.deepseek.com/v1` |
 | OpenRouter | `https://openrouter.ai/api` | `https://openrouter.ai/api/v1` |
 
 To switch to OpenAI-compatible endpoints, modify the vendor's `baseUrl` and `defaultApiStyle`.
@@ -193,7 +193,7 @@ Limited by VS Code's public API, this extension additionally implements context
 
 ### Intelligent Commit Message Generation
 
-1. Press `Ctrl+Shift+P`, type `编码套餐: 生成 Commit 消息`
+1. Press `Ctrl+Shift+P`, type `Coding Plans: Generate Commit Message`
 2. The extension analyzes current Git changes and automatically generates a Conventional Commits-compliant commit message
 3. You can select the model to use (defaults to the currently configured vendor)
 

package.json

@@ -434,15 +434,15 @@
                             "defaultApiStyle": "anthropic"
                         },
                         {
-                            "name": "Infini",
-                            "baseUrl": "https://cloud.infini-ai.com",
+                            "name": "Qiniu",
+                            "baseUrl": "https://api.qnaigc.com",
                             "models": [],
                             "useModelsEndpoint": true,
                             "defaultApiStyle": "anthropic"
                         },
                         {
-                            "name": "Qiniu",
-                            "baseUrl": "https://api.qnaigc.com",
+                            "name": "DeepSeek",
+                            "baseUrl": "https://api.deepseek.com/anthropic",
                             "models": [],
                             "useModelsEndpoint": true,
                             "defaultApiStyle": "anthropic"

src/config/configStore.ts

@@ -433,7 +433,11 @@ export class ConfigStore implements vscode.Disposable {
         normalized.push(this.buildStoredModelEntry({
           ...inputModel,
           temperature: undefined,
-          topP: undefined
+          topP: undefined,
+          capabilities: {
+            ...inputModel.capabilities,
+            vision: defaultVision
+          }
         }, canonical, defaultVision, defaultApiStyle));
         continue;
       }

src/providers/genericProviderDiscovery.ts

@@ -67,10 +67,6 @@ export function mergeConfiguredModelOverrides(
   return discoveredModels.map(discovered => {
     const configured = configuredByName.get(discovered.name.trim().toLowerCase());
     if (!configured) {
-      const discoveredVision = discovered.capabilities?.vision;
-      if (typeof discoveredVision === 'boolean') {
-        return discovered;
-      }
       return {
         ...discovered,
         capabilities: {

src/test/runTest.ts

@@ -426,7 +426,7 @@ const testCases: TestCase[] = [
     }
   },
   {
-    name: '新增模型写回时保留必要发现字段但不落 token 上限',
+    name: '新增模型写回时使用供应商 defaultVision 且不落 token 上限',
     initialVendors: [createVendorWithSpacedModelName()],
     discoveredModels: [
       { name: 'gpt-4o' },
@@ -448,7 +448,7 @@ const testCases: TestCase[] = [
       assert.equal(newModel?.description, 'Fresh from /models');
       assert.equal(newModel?.temperature, undefined);
       assert.equal(newModel?.topP, undefined);
-      assert.deepEqual(newModel?.capabilities, { tools: true, vision: true });
+      assert.deepEqual(newModel?.capabilities, { tools: true, vision: false });
       assert.equal(newModel?.maxInputTokens, undefined);
       assert.equal(newModel?.maxOutputTokens, undefined);
     }
@@ -1121,6 +1121,72 @@ function runGenericProviderContextSizeTests(
   }
 }
 
+async function runGenericProviderDiscoveryDefaultVisionTests(
+  configStoreCtor: ConfigStoreCtor,
+  genericProviderModule: GenericProviderModule
+): Promise<void> {
+  const { GenericAIProvider } = genericProviderModule;
+  const originalFetch = globalThis.fetch;
+
+  activeState = createState([{
+    name: 'Vendor',
+    baseUrl: 'https://example.test/v1',
+    defaultApiStyle: 'openai-chat',
+    defaultVision: false,
+    useModelsEndpoint: true,
+    models: []
+  }]);
+
+  const configStore = new configStoreCtor(createExtensionContext() as never);
+  const provider = new GenericAIProvider(createExtensionContext() as never, configStore) as unknown as {
+    models: Array<{
+      id: string;
+      capabilities?: {
+        imageInput?: boolean;
+        toolCalling?: boolean | number;
+      };
+    }>;
+    refreshModels(): Promise<void>;
+    dispose(): void;
+  };
+
+  globalThis.fetch = (async (_url: string | URL | Request): Promise<Response> => {
+    return new Response(JSON.stringify({
+      data: [{
+        id: 'fresh-vision-model',
+        context_length: 64000,
+        capabilities: {
+          tool_calling: true,
+          image_input: true
+        }
+      }]
+    }), {
+      status: 200,
+      headers: {
+        'content-type': 'application/json'
+      }
+    });
+  }) as typeof globalThis.fetch;
+
+  try {
+    (configStore as unknown as { getApiKey(vendorName: string): Promise<string> }).getApiKey = async (vendorName: string) => (
+      vendorName === 'Vendor' ? 'configured' : ''
+    );
+
+    await provider.refreshModels();
+
+    const updatedVendor = getUpdatedVendor(activeState);
+    const refreshedModel = updatedVendor.models.find(model => model.name === 'fresh-vision-model');
+    assert.deepEqual(refreshedModel?.capabilities, { tools: true, vision: false });
+    assert.equal(provider.models[0]?.capabilities?.imageInput, false);
+    console.log('PASS /models 刷新新增模型时 defaultVision=false 会覆盖发现到的 vision=true');
+  } finally {
+    globalThis.fetch = originalFetch;
+    provider.dispose();
+    configStore.dispose();
+  }
+}
+
 async function runGenericProviderEmptyResponseTests(
   configStoreCtor: ConfigStoreCtor,
   genericProviderModule: GenericProviderModule
@@ -3207,6 +3273,7 @@ async function main(): Promise<void> {
     await runConfigNormalizationTests(ConfigStore);
     runTokenWindowResolutionTests(baseProviderModule);
     runGenericProviderContextSizeTests(ConfigStore, genericProviderModule);
+    await runGenericProviderDiscoveryDefaultVisionTests(ConfigStore, genericProviderModule);
     await runGenericProviderEmptyResponseTests(ConfigStore, genericProviderModule);
     await runGenericProviderOutputLimitToggleTests(ConfigStore, genericProviderModule, tokenUsageModule);
     await runGenericProviderAnthropicSamplingCompatibilityTests(ConfigStore, genericProviderModule);