PageSourceSearch

https://cdn.wisgate.ai/wisgate/_nuxt/B10wNirm.js

js wisgate.ai collected 2026-10-04 13:26:27 UTC 83,559 bytes, 2,404 lines download raw bytes

1import{d as y,u as A,g as I,a as r,al as v,U as _,j as w,s as b,o as P,l as e,k as c,n as G,p,m as u,aM as k,W}from"./DZ7mik94.js";import{_ as S}from"./56p-yg89.js";import{i as O,g as L}from"./Dog5d0bc.js";import"./iik6CYzq.js";const E=`# Claude Code IDE Plugin
2
3Claude Code is an intelligent coding tool that can run in the terminal or be used as a plugin in VSCode, JetBrains, and other IDEs.
4
5## Step 1: Install Claude Code & Configure WisGate
6
7Please first follow the Claude Code guide to complete the Claude Code installation and WisGate configuration.
8
9## Step 2: Install IDE Plugin
10
11### Claude Code for VS Code
12
13The Claude Code VS Code extension works in VS Code and VS Code-based IDEs (such as Cursor, Trae, etc.).
14
151. Open VS Code, go to the Extensions marketplace, search for "Claude Code" and install it
162. After installation, click the Claude Code icon in the top right corner
173. Wait a few seconds for initialization, then you can start using it
184. Enter \`/config\` in the Claude Code dialog, and check \`Disable Login Prompt\` to close the login page
19
20### Claude Code for JetBrains
21
22The Claude Code JetBrains plugin supports JetBrains IDEs such as IntelliJ IDEA, PyCharm, WebStorm, etc.
23
241. Open your JetBrains IDE, go to the plugin marketplace, search for "Claude Code" and install it
252. After installation, restart the IDE to start using Claude Code
26
27Done! You can now use Claude Code in your IDE for development.
28`,T=`# Claude Code IDE 插件
29
30Claude Code 是一个智能编码工具,可以在终端中运行,也可以通过在 VSCode、JetBrains 等 IDE 安装插件使用。
31
32## 步骤一:安装 Claude Code & 配置 WisGate
33
34请先参考 Claude Code 指南完成 Claude Code 的安装与 WisGate 的配置。
35
36## 步骤二:安装使用 IDE 插件
37
38### Claude Code for VS Code 插件
39
40Claude Code VSCode 插件支持在 VSCode 及基于 VSCode 的 IDE(如 Cursor、Trae 等)中使用。
41
421. 打开 VSCode,进入扩展市场,搜索并安装「Claude Code」插件
432. 安装完成后,点击右上角 Claude Code 图标,即进入 Claude Code 页面
443. 在登录页面等待几秒待其初始化完成后即可使用
454. 成功进入 Claude Code 对话框后,在对话框输入 \`/config\` 进入设置,勾选 \`Disable Login Prompt\` 配置来关闭登录页面
46
47### Claude Code for JetBrains 插件
48
49Claude Code JetBrains 插件支持 JetBrains 的系列 IDE 如 IntelliJ IDEA、PyCharm、WebStorm 等。
50
511. 打开 JetBrains IDE,进入插件市场,搜索「Claude Code」插件并进行安装
522. 安装完成后,重启 IDE 即可使用 Claude Code 功能
53
54完毕!现在就可以在 IDE 中正常使用 Claude Code 进行开发了。
55`,U=`# Claude Code
56
57Claude Code is an intelligent coding agent that runs in the terminal. By connecting Claude Code to WisGate, developers can use WisGate's Anthropic-compatible endpoint for code generation, debugging, refactoring, project analysis, and automated development workflows.
58
59## Step 1: Install Claude Code
60
61Prerequisites:
62- Node.js 18 or later
63- MacOS users are recommended to use nvm to install Node.js
64- Windows users also need Git for Windows
65
66Install Claude Code:
67
68\`\`\`bash
69npm install -g @anthropic-ai/claude-code
70\`\`\`
71
72Verify installation:
73
74\`\`\`bash
75claude --version
76\`\`\`
77
78## Step 2: Configure WisGate
79
80### 1. Sign Up
81
82Visit the WisGate platform and complete the registration process.
83
84### 2. Get Your API Key
85
86Create and obtain your API Key from the WisGate dashboard.
87
88### 3. Configure Claude Code Environment Variables
89
90Edit or create the \`settings.json\` file:
91- MacOS & Linux: \`~/.claude/settings.json\`
92- Windows: \`User Directory/.claude/settings.json\`
93
94\`\`\`json
95{
96  "env": {
97    "ANTHROPIC_AUTH_TOKEN": "your_wisgate_api_key",
98    "ANTHROPIC_API_KEY": "",
99    "ANTHROPIC_BASE_URL": "https://api.wisgate.ai",
100    "API_TIMEOUT_MS": "3000000",
101    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": 1
102  }
103}
104\`\`\`
105
106Use \`https://api.wisgate.ai\` as the WisGate Anthropic-compatible endpoint. Do not use the OpenAI-compatible endpoint \`https://api.wisgate.ai/v1\` for Claude Code. Setting \`ANTHROPIC_API_KEY\` to an empty string helps prevent Claude Code from picking up an old Anthropic API key.
107
108Edit or create \`.claude.json\`:
109- MacOS & Linux: \`~/.claude.json\`
110- Windows: \`User Directory/.claude.json\`
111
112\`\`\`json
113{
114  "hasCompletedOnboarding": true
115}
116\`\`\`
117
118If you previously logged in to Claude Code with an Anthropic account, run \`/logout\` once inside Claude Code, then quit the session and reopen your terminal. This prevents cached login state from conflicting with your WisGate API key.
119
120## Step 3: Start Using Claude Code
121
122Navigate to your code directory and run \`claude\` to start using Claude Code.
123
124If prompted with "Do you want to use this API key", select Yes. After entering Claude Code, run \`/status\` to verify that the auth token is \`ANTHROPIC_AUTH_TOKEN\` and the base URL is \`https://api.wisgate.ai\`.
125
126## WisGate Use Cases
127
128- **Code generation**: Describe a feature and let Claude Code generate functions, pages, scripts, or configuration files through WisGate.
129- **Debugging**: Ask Claude Code to inspect errors, identify causes, and use WisGate models to help fix issues.
130- **Project refactoring**: Split modules, improve naming, organize dependencies, and reduce duplicate logic in larger codebase
130s.
131- **Test creation**: Generate unit tests, integration tests, and edge cases from existing business logic.
132- **Team access management**: Standardize WisGate API key and model configuration across team members for easier model and cost management.
133
134## FAQ
135
136### How to Switch Models
137
138Adjust the configuration file (\`~/.claude/settings.json\`) to assign different WisGate models for Claude Code tasks:
139
140\`\`\`json
141{
142  "env": {
143    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "claude-haiku-4-5-20251001",
144    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-4-6",
145    "ANTHROPIC_DEFAULT_OPUS_MODEL": "claude-opus-4-8"
146  }
147}
148\`\`\`
149
150Claude Code has the best compatibility with Anthropic-family models. Prefer Claude models on WisGate that support the Anthropic Messages format.
151
152### How to Verify the WisGate Connection
153
154Run this command inside Claude Code:
155
156\`\`\`bash
157/status
158\`\`\`
159
160If the status shows \`ANTHROPIC_AUTH_TOKEN\` and \`https://api.wisgate.ai\`, Claude Code is connected through WisGate.
161
162### Configuration Changes Not Taking Effect
163
164- Close all Claude Code windows, reopen a new terminal, and run \`claude\` again
165- If you previously logged in with an Anthropic account, run \`/logout\` to clear cached login state
166- Make sure there is no real \`ANTHROPIC_API_KEY\` left in your shell environment, or set it to an empty string in \`settings.json\`
167- Verify the JSON format is correct
168
169### Recommended Claude Code Version
170
171\`\`\`bash
172# Check current version
173claude --version
174
175# Upgrade to latest version
176claude update
177\`\`\`
178`,x=`# Claude Code
179
180Claude Code 是一个在终端中运行的智能编程代理。通过接入 WisGate,开发者可以在 Claude Code 中使用 WisGate 提供的 Anthropic 兼容接口完成代码生成、调试、重构、项目分析和自动化开发任务。
181
182## 步骤一:安装 Claude Code
183
184前提条件:
185- 您需要安装 Node.js 18 或更新版本环境
186- MacOS 用户推荐使用 nvm 方式安装 Nodejs
187- Windows 用户还需安装 Git for Windows
188
189进入命令行界面,安装 Claude Code:
190
191\`\`\`bash
192npm install -g @anthropic-ai/claude-code
193\`\`\`
194
195运行如下命令,查看安装结果,若显示版本号则表示安装成功:
196
197\`\`\`bash
198claude --version
199\`\`\`
200
201## 步骤二:配置 WisGate
202
203### 1. 注册账号
204
205访问 WisGate 开放平台,完成账号注册流程。
206
207### 2. 获取 API Key
208
209在 WisGate 控制台中创建并获取您的 API Key。
210
211### 3. 配置 Claude Code 环境变量
212
213编辑或新增 \`settings.json\` 文件:
214- MacOS & Linux 为 \`~/.claude/settings.json\`
215- Windows 为 \`用户目录/.claude/settings.json\`
216
217\`\`\`json
218{
219  "env": {
220    "ANTHROPIC_AUTH_TOKEN": "your_wisgate_api_key",
221    "ANTHROPIC_API_KEY": "",
222    "ANTHROPIC_BASE_URL": "https://api.wisgate.ai",
223    "API_TIMEOUT_MS": "3000000",
224    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": 1
225  }
226}
227\`\`\`
228
229其中 \`ANTHROPIC_BASE_URL\` 使用 WisGate 的 Anthropic 兼容入口 \`https://api.wisgate.ai\`,不要填写 OpenAI 兼容入口 \`https://api.wisgate.ai/v1\`。\`ANTHROPIC_API_KEY\` 建议显式置空,避免 Claude Code 继续读取旧的 Anthropic Key。
230
231再编辑或新增 \`.claude.json\` 文件:
232- MacOS & Linux 为 \`~/.claude.json\`
233- Windows 为 \`用户目录/.claude.json\`
234
235\`\`\`json
236{
237  "hasCompletedOnboarding": true
238}
239\`\`\`
240
241如果您之前已经用 Anthropic 账号登录过 Claude Code,请先在 Claude Code 中执行一次 \`/logout\`,然后关闭当前会话并重新打开终端,避免缓存登录态与 WisGate API Key 冲突。
242
243## 步骤三:开始使用 Claude Code
244
245配置完成后,进入一个您的代码工作目录,在终端中执行 \`claude\` 命令即可开始使用。
246
247若遇到「Do you want to use this API key」选择 Yes 即可。进入 Claude Code 后,可以执行 \`/status\` 检查当前连接状态,确认 Auth token 使用 \`ANTHROPIC_AUTH_TOKEN\`,Base URL 为 \`https://api.wisgate.ai\`。
248
249## WisGate 使用场景
250
251- **代码生成**:在 Claude Code 中描述需求,让模型生成函数、页面、脚本或配置文件。
252- **代码调试**:让 Claude Code 读取报错、定位原因,并通过 WisGate 模型辅助修复问题。
253- **项目重构**:在大型代码库中拆分模块、优化命名、整理依赖和减少重复逻辑。
254- **测试补全**:根据现有业务逻辑生成单元测试、集成测试或边界用例。
255- **团队统一接入**:团队成员统一使用 WisGate API Key 和模型配置,便于后续管理模型调用和成本。
256
257## 常见问题
258
259### 如何切换使用模型
260
261直接通过调整配置文件的方式(位于 \`~/.claude/settings.json\`)来指定不同任务使用的 WisGate 模型:
262
263\`\`\`json
264{
265  "env": {
266    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "claude-haiku-4-5-20251001",
267    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-4-6",
268    "ANTHROPIC_DEFAULT_OPUS_MODEL": "claude-opus-4-8"
269  }
270}
271\`\`\`
272
273Claude Code 对 Anthropic 系列模型兼容性最好。建议优先选择 WisGate 上支持 Anthropic Messages 格式的 Claude 模型。
274
275### 如何验证 WisGate 是否接入成功
276
277在 Claude Code 内执行:
278
279\`\`\`bash
280/status
281\`\`\`
282
283若状态中显示 \`ANTHROPIC_AUTH_TOKEN\` 和 \`https://api.wisgate.ai\`,说明 Claude Code 已经通过 WisGate 连接。
284
285### 手工修改配置不生效
286
287- 关闭所有 Claude Code 窗口,重新打开一个新的命令行窗口,再次运行 \`claude\` 启动
288- 如果之前登录过 Anthropic 账号,请先执行 \`/logout\` 清理缓存登录态
289- 确认系统环境中没有残留真实的 \`ANTHROPIC_API_KEY\`,或已在 \`settings.json\` 中设置为空字符串
290- 确认配置文件的 JSON 格式是否正确
291
292### 推荐的 Claude Code 版本
293
294\`\`\`bash
295# 检查当前版本
296claude --version
297
298# 升级到最新版本
299claude update
300\`\`\`
301`,M=`# Claude Desktop
302
303Claude Desktop supports third-party inference gateways. For Anthropic Messages-compatible models, you can connect Claude Desktop to WisGate through the Gateway configuration.
304
305## 1. Install Claude Desktop
306
307Download and install Claude Desktop from Anthropic. Start the app after installation.
308
309## 2. Enable Developer Mode
310
3111. Open Claude Desktop
3122. Go to **Help > Troubleshooting**
3133. Click **Enable Developer Mode**
3144. After enabling it, a **Developer** menu appears in the menu bar
315
316## 3. Configure Third-Party Inference
317
318Open **Developer > Configure Third-Party Inference...**, choose Gateway / Anthropic-compatible mode, and enter:
319
320- **Gateway base URL**: \`https://api.wisgate.ai\`
321- **Gateway API key**: Your WisGate API key
322- **Gateway auth scheme**: \`bearer\`
323
324If the interface asks for the full Messages API URL instead of a Base URL, enter:
325
326\`\`\`text
327https://api.wisgate.ai/v1/messages
328\`\`\`
329
330Save the settings, fully quit Claude Desktop, then reopen it.
331
332## 4. Select a Model and Start Using
333
334After restart, choose the Gateway or local configuration entry on the start screen. Inside Claude Desktop, select an available Anthropic-compatible model from the model picker.
335
336## Use Cases
337
338- **Desktop chat**: Use Claude-family or Anthropic Messages-compatible models through WisGate in Claude Desktop.
339- **Unified team gateway**: Standardize API key and model access through one gateway.
340- **MCP tools**: Combine WisGate model access with Claude Desktop's MCP tool capabilities.
341
342## FAQ
343
344### The Gateway entry does not appear
345
346Confirm that Developer Mode is enabled, then fully quit and reopen Claude Desktop.
347
348### Connection fails
349
350Confirm that the Base URL is \`https://api.wisgate.ai\`, auth scheme is \`bearer\`, and the API key is correct. If the interface asks for a full URL, use \`https://api.wisgate.ai/v1/messages\`.
351`,R=`# Claude Desktop
352
353Claude Desktop 支持第三方推理网关。对于 Anthropic Messages 兼容模型,可以通过 Gateway 配置接入 WisGate。
354
355## 一、安装 Claude Desktop
356
357访问 Anthropic 官网下载并安装 Claude Desktop。安装完成后启动应用。
358
359## 二、启用 Developer Mode
360
3611. 打开 Claude Desktop
3622. 进入 **Help > Troubleshooting**
3633. 点击 **Enable Developer Mode**
3644. 启用后,菜单栏会出现 **Developer** 菜单
365
366## 三、配置第三方推理网关
367
368进入 **Developer > Configure Third-Party Inference...**,选择 Gateway / Anthropic-compatible 后填写:
369
370- **Gateway base URL**:\`https://api.wisgate.ai\`
371- **Gateway API key**:您的 WisGate API Key
372- **Gateway auth scheme**:\`bearer\`
373
374如果界面要求填写完整 Messages API 地址,而不是 Base URL,请填写:
375
376\`\`\`text
377https://api.wisgate.ai/v1/messages
378\`\`\`
379
380保存后完全退出 Claude Desktop,再重新打开。
381
382## 四、选择模型并使用
383
384重新启动后,在启动页选择 Gateway 或本地配置入口。进入 Claude Desktop 后,在模型选择器中选择可用的 Anthropic 兼容模型。
385
386## 使用场景
387
388- **桌面端对话**:在 Claude Desktop 中通过 WisGate 使用 Claude 系列或 Anthropic Messages 兼容模型。
389- **团队统一网关**:统一 API Key 和模型访问入口。
390- **MCP 工具使用**:结合 Claude Desktop 的 MCP 能力连接本地工具和外部服务。
391
392## 常见问题
393
394### Gateway 入口没有出现
395
396请确认已经启用 Developer Mode,并完全退出后重新打开 Claude Desktop。
397
398### 连接失败
399
400请确认 Base URL 为 \`https://api.wisgate.ai\`,auth scheme 为 \`bearer\`,API Key 正确。如果界面要求完整地址,请使用 \`https://api.wisgate.ai/v1/messages\`。
401`,j=`# Cline
402
403Cline is a powerful VS Code extension that helps you use AI models directly in the editor for code generation, file operations, and more. It not only speeds up development but also provides contextual understanding and intelligent suggestions for c
403omplex tasks.
404
405When using WisGate, configure the API endpoint \`https://api.wisgate.ai/v1\`.
406
407## 1. Install Cline Extension
408
409### 1. Open Extension Marketplace
410
4111. Open VS Code
4122. Click the extension marketplace icon on the left
4133. Search for \`cline\` in the search box
4144. Find the \`Cline\` extension
415
416### 2. Install Extension
417
4181. Click the \`Install\` button
4192. After installation, select trust developer
420
421## 2. Configure API Settings
422
423### 1. Select API Key Method
424
425Select \`Use your own API Key\`
426
427### 2. Enter Configuration
428
429- **API Provider**: Select \`OpenAI Compatible\`
430- **Base URL**: Enter \`https://api.wisgate.ai/v1\`
431- **API Key**: Enter your WisGate API Key
432- **Model**: Select "Use custom" and enter the model name (e.g., \`gpt-5.5\`)
433- **Other settings**:
434  - Uncheck **Support Images**
435  - Set **Context Window Size** to \`200000\`
436  - Adjust \`temperature\` and other parameters as needed
437
438## 3. Start Using
439
440After configuration, you can enter your requirements in the input box and let the model help you with various tasks:
441
442- Create and edit files
443- Generate code
444- Refactor code
445- Explain code logic
446- Debug issues
447
448## 4. Usage Tips
449
450- For tasks that need file reads or writes, confirm Cline's file access permissions first.
451- For large projects, specify the target folder, target files, and expected change scope.
452- If your task does not need vision capability, keep **Support Images** disabled to avoid unnecessary compatibility issues.
453
454## FAQ
455
456### API requests fail
457
458Confirm that **API Provider** is \`OpenAI Compatible\`, **Base URL** is \`https://api.wisgate.ai/v1\`, and the API key has no extra spaces.
459
460### Context length is not enough
461
462Increase **Context Window Size**, or split the task into smaller steps and let Cline handle them in stages.
463`,K=`# Cline
464
465Cline 是一个强大的 VS Code 插件,可以帮助您在编辑器中直接使用 AI 模型进行代码生成、文件操作等任务。它不仅能加快开发节奏,还能在复杂任务中提供上下文理解和智能建议,让开发过程更高效、更流畅。
466
467使用 WisGate 时,配置 API 端点 \`https://api.wisgate.ai/v1\` 即可。
468
469## 一、安装 Cline 插件
470
471### 1. 打开插件市场
472
4731. 打开 VS Code
4742. 点击左侧插件市场图标
4753. 在搜索框中输入 \`cline\`
4764. 找到 \`Cline\` 扩展
477
478### 2. 安装插件
479
4801. 点击 \`Install\` 按钮进行安装
4812. 安装完成后,选择信任开发者
482
483## 二、配置 API 设置
484
485### 1. 选择 API Key 方式
486
487选择 \`Use your own API Key\`
488
489### 2. 填入配置信息
490
491请按照以下配置填入相关信息:
492
493- **API Provider**:选择 \`OpenAI Compatible\`
494- **Base URL**:输入 \`https://api.wisgate.ai/v1\`
495- **API Key**:填入您的 WisGate API Key
496- **模型**:选择"使用自定义",并输入模型名称(如:\`gpt-5.5\`)
497- **其他配置**:
498  - 取消勾选 **Support Images**
499  - 调整 **Context Window Size** 为 \`200000\`
500  - 根据您的任务需求调整 \`temperature\` 等其它参数
501
502## 三、开始使用
503
504配置完成后,您可以在输入框中输入需求,让模型帮助您完成各种任务,例如:
505
506- 创建和编辑文件
507- 生成代码
508- 重构代码
509- 解释代码逻辑
510- 调试问题
511
512## 四、使用建议
513
514- 对需要读写文件的任务,先确认 Cline 的文件访问权限。
515- 对长上下文项目,优先说明目标目录、目标文件和预期修改范围。
516- 如果模型不需要视觉能力,保持 **Support Images** 关闭,减少不必要的兼容性问题。
517
518## 常见问题
519
520### API 请求失败
521
522请确认 **API Provider** 为 \`OpenAI Compatible\`,**Base URL** 为 \`https://api.wisgate.ai/v1\`,API Key 没有多余空格。
523
524### 上下文长度不足
525
526可以调高 **Context Window Size**,或将任务拆分成多个较小步骤,让 Cline 分阶段处理。
527`,D=`# CodeBuddy
528
529CodeBuddy is an AI-powered full-process programming tool. It can connect to WisGate through local model configuration so product, design, development, and deployment workflows use a unified model service.
530
531## 1. Install CodeBuddy
532
533Visit the CodeBuddy website to download and install the version for your operating system.
534
535## 2. Log In to CodeBuddy
536
537After downloading and installing, log in to CodeBuddy as prompted.
538
539## 3. Configure WisGate
540
541### 1. Configuration File Location
542
543\`models.json\` is a configuration file for customizing the model list and controlling the model dropdown display. It is recommended to configure it at the user level: \`~/.codebuddy/models.json\`.
544
545### 2. Configuration Structure
546
547\`\`\`json
548{
549    "models": [
550        {
551            "id": "gpt-5.5",
552            "name": "WisGate GPT",
553            "vendor": "wisgate",
554            "apiKey": "your_wisgate_api_key",
555            "url": "https://api.wisgate.ai/v1",
556            "supportsToolCall": true,
557            "supportsReasoning": true
558        }
559    ]
560}
561\`\`\`
562
563### 3. Start Using
564
565Select the configured model in the chat and start a conversation.
566
567## 4. Use Cases
568
569- **Requirement to code**: Generate APIs, pages, or business logic from product descriptions.
570- **Code explanation**: Ask the model to explain modules, call chains, and key logic.
571- **Debugging**: Provide error logs and related files to help locate issues.
572- **Development collaboration**: Standardize model calls across the team with the same WisGate configuration.
573
574## FAQ
575
576### The new model does not appear in the dropdown
577
578Confirm that \`models.json\` is saved at \`~/.codebuddy/models.json\`, the JSON format is valid, and restart CodeBuddy.
579
580### Requests fail or authentication fails
581
582Check that \`apiKey\` is a key created in the WisGate dashboard and \`url\` is \`https://api.wisgate.ai/v1\`.
583`,B=`# CodeBuddy
584
585CodeBuddy 是一款基于 AI 的全流程智能编程工具,支持通过本地模型配置接入 WisGate,让产品、设计、研发和部署流程中的编码任务使用统一模型服务。
586
587## 一、安装 CodeBuddy
588
589访问 CodeBuddy 官网下载并安装适合您的操作系统的版本。
590
591## 二、登陆 CodeBuddy
592
593下载安装后根据提示登录 CodeBuddy。
594
595## 三、配置 WisGate
596
597### 1. 配置文件位置
598
599\`models.json\` 是一个配置文件,用于自定义模型列表和控制模型下拉列表的显示。建议在用户级配置该文件:\`~/.codebuddy/models.json\`。
600
601### 2. 配置结构
602
603\`\`\`json
604{
605    "models": [
606        {
607            "id": "gpt-5.5",
608            "name": "WisGate GPT",
609            "vendor": "wisgate",
610            "apiKey": "your_wisgate_api_key",
611            "url": "https://api.wisgate.ai/v1",
612            "supportsToolCall": true,
613            "supportsReasoning": true
614        }
615    ]
616}
617\`\`\`
618
619### 3. 回到 CodeBuddy 进行对话
620
621在对话框选择配置好的模型进行对话即可。
622
623## 四、使用场景
624
625- **需求到代码**:根据产品描述生成接口、页面或业务逻辑。
626- **代码解释**:让模型解释现有模块、调用链和关键逻辑。
627- **调试排错**:将错误日志和相关文件交给模型辅助定位。
628- **研发协作**:在同一套 WisGate 配置下统一团队模型调用方式。
629
630## 常见问题
631
632### 模型下拉列表没有出现新模型
633
634请确认 \`models.json\` 放在 \`~/.codebuddy/models.json\`,JSON 格式正确,并重启 CodeBuddy。
635
636### 请求失败或鉴权失败
637
638请检查 \`apiKey\` 是否为 WisGate 控制台创建的 Key,\`url\` 是否为 \`https://api.wisgate.ai/v1\`。
639`,N=`# Codex CLI
640
641Codex CLI is OpenAI's local terminal coding agent. It supports custom model providers through OpenAI-compatible endpoints. After connecting Codex CLI to WisGate, you can use WisGate models for code generation, project analysis, file edits, and debugging tasks in the terminal.
642
643## 1. Install Codex CLI
644
645Follow the official Codex CLI installation instructions first. After installation, verify that the command is available:
646
647\`\`\`bash
648codex --version
649\`\`\`
650
651## 2. Configure WisGate
652
653Codex CLI usually reads \`~/.codex/config.toml\`. Create the file if it does not exist.
654
655\`\`\`toml
656model_provider = "wisgate"
657model = "gpt-5.5"
658model_reasoning_effort = "high"
659
660[model_providers.wisgate]
661name = "WisGate"
662base_url = "https://api.wisgate.ai/v1"
663env_key = "WISGATE_API_KEY"
664\`\`\`
665
666Then set your WisGate API key in the shell:
667
668\`\`\`bash
669export WISGATE_API_KEY="your_wisgate_api_key"
670\`\`\`
671
672For persistent configuration, add the export line to \`~/.zshrc\`, \`~/.bashrc\`, or your shell profile.
673
674## 3. Start Using
675
676Navigate to your project directory and start Codex CLI:
677
678\`\`\`bash
679cd /path/to/your/project
680codex
681\`\`\`
682
683Codex CLI will call the configured model through \`https://api.wisgate.ai/v1\`.
684
685## Use Cases
686
687- **Terminal development**: Analyze projects, generate code, and modify files from the command line.
688- **Code review**: Ask Codex CLI to inspect changes, summarize risks, and suggest fixes.
689- **Refactoring tasks**: Split complex modules, adjust structure, and run verification commands.
690- **Team configuration**: Standardize WisGate access with a shared \`config.toml\` template.
691
692## FAQ
693
694### Authentication fails
695
696Confirm that \`WISGATE_API_KEY\` is set and that \`env_key\` matches the environment variable name.
697
698### The model is not available
699
700Confirm that \`model\` is a model ID supported by WisGate. The OpenAI-compatible endpoint must be \`https://api.wisgate.ai/v1\`.
701`,F=`# Codex CLI
702
703Codex CLI 是 OpenAI 的本地终端编程代理,支持通过自定义模型服务商接入 OpenAI 兼容接口。接入 WisGate 后,可以在终端中使用 WisGate 模型完成代码生成、项目分析、文件修改和调试任务。
704
705## 一、安装 Codex CLI
706
707请先按照 Codex CLI 官方说明完成安装。安装完成后,运行以下命令确认可用:
708
709\`\`\`bash
710codex --version
711\`\`\`
712
713## 二、配置 WisGate
714
715Codex CLI 通常读取 \`~/.codex/config.toml\`。如果文件不存在,请手动创建。
716
717\`\`\`toml
718model_provider = "wisgate"
719model = "gpt-5.5"
720model_reasoning_effort = "high"
721
722[model_providers.wisgate]
723name = "WisGate"
724base_url = "https://api.wisgate.ai/v1"
725env_key = "WISGATE_API_KEY"
726\`\`\`
727
728然后在 shell 中设置 WisGate API Key:
729
730\`\`\`bash
731export WISGATE_API_KEY="your_wisgate_api_key"
732\`\`\`
733
734如果希望长期生效,请把这行写入 \`~/.zshrc\`、\`~/.bashrc\` 或对应 shell 配置文件。
735
736## 三、开始使用
737
738进入项目目录后启动 Codex CLI:
739
740\`\`\`bash
741cd /path/to/your/project
742codex
743\`\`\`
744
745启动后,Codex CLI 会通过 \`https://api.wisgate.ai/v1\` 调用所配置的模型。
746
747## 使用场景
748
749- **终端内开发**:在命令行中分析项目、生成代码和修改文件。
750- **代码审查**:让 Codex CLI 检查变更、总结风险并给出修复建议。
751- **重构任务**:拆分复杂模块、调整结构并运行验证命令。
752- **团队统一配置**:通过同一套 \`config.toml\` 模板统一接入 WisGate。
753
754## 常见问题
755
756### 鉴权失败
757
758请确认 \`WISGATE_API_KEY\` 已设置,且 \`env_key\` 与环境变量名称一致。
759
760### 模型不可用
761
762请确认 \`model\` 填写的是 WisGate 支持的模型 ID。OpenAI 兼容入口必须使用 \`https://api.wisgate.ai/v1\`。
763`,H=`# Crush
764
765Crush is a CLI + TUI AI programming tool that runs in the terminal, supporting multiple model integrations for code generation, debugging, conversation, file operations, and multitasking.
766
767When using WisGate, configure the API endpoint \`https://api.wisgate.ai/v1\`.
768
769## 1. Install Crush
770
771Choose the installation method for your system:
772
773**Homebrew (macOS recommended):**
774
775\`\`\`bash
776brew install charmbracelet/tap/crush
777\`\`\`
778
779**NPM (cross-platform):**
780
781\`\`\`bash
782npm install -g @charmland/crush
783\`\`\`
784
785## 2. Configure WisGate Models
786
787### 1. Get Your API Key
788
789Create and obtain your API Key from the WisGate dashboard.
790
791### 2. Start Crush and Select Model
792
793\`\`\`bash
794crush
795\`\`\`
796
797In the model selection screen, select your desired model (e.g., \`gpt-5.5\`)
798
799### 3. Enter API Key
800
801Enter your WisGate API Key when prompted.
802
803### 4. Modify Crush Configuration
804
805Configuration file location:
806- macOS/Linux: \`~/.config/crush/crush.json\`
807- Windows: \`%USERPROFILE%\\.config\\crush\\crush.json\`
808
809\`\`\`json
810{
811  "providers": {
812    "wisgate": {
813      "id": "wisgate",
814      "name": "WisGate Provider",
815      "base_url": "https://api.wisgate.ai/v1",
816      "api_key": "your_wisgate_api_key"
817    }
818  }
819}
820\`\`\`
821
822After configuration, restart Crush and use \`/models\` to select a specific model.
823
824## 3. Use Cases
825
826- **Terminal coding**: Run the TUI inside a project directory for code generation and edits.
827- **File operations**: Let Crush read project context and assist with file edits.
828- **Debugging**: Use command output, logs, and test results to generate fix suggestions.
829- **Multitasking**: Split complex development requests into terminal-based tasks and complete them step by step.
830
831## FAQ
832
833### The configured model does not appear in \`/models\`
834
835Confirm that the configuration file path is correct, the JSON format is valid, and restart Crush.
836
837### API calls fail
838
839Confirm that \`base_url\` is \`https://api.wisgate.ai/v1\` and \`api_key\` has been replaced with an API key created in the WisGate dashboard.
840`,z=`# Crush
841
842Crush 既是一款在终端中运行的 CLI + TUI AI 编程工具,也支持多种模型接入,能够在命令行环境下完成代码生成、调试、对话、文件操作与多任务处理等工作。
843
844使用 WisGate 时,配置 API 端点 \`https://api.wisgate.ai/v1\` 即可。
845
846## 一、安装 Crush
847
848根据您的系统选择对应的安装方式:
849
850**Homebrew(macOS 推荐):**
851
852\`\`\`bash
853brew install charmbracelet/tap/crush
854\`\`\`
855
856**NPM(跨平台):**
857
858\`\`\`bash
859npm install -g @charmland/crush
860\`\`\`
861
862## 二、配置 WisGate 模型
863
864### 1. 获取 API 密钥
865
866在 WisGate 控制台中创建并获取您的 API Key。
867
868### 2. 启动 Crush 并选择模型
869
870\`\`\`bash
871crush
872\`\`\`
873
874在模型选择界面中,选择您需要的模型(如:\`gpt-5.5\`)
875
876### 3. 输入 API 密钥
877
878在提示界面中输入您的 WisGate API Key。
879
880### 4. 修改 Crush 配置
881
882配置文件位置:
883- macOS/Linux:\`~/.config/crush/crush.json\`
884- Windows:\`%USERPROFILE%\\.config\\crush\\crush.json\`
885
886\`\`\`json
887{
888  "providers": {
889    "wisgate": {
890      "id": "wisgate",
891      "name": "WisGate Provider",
892      "base_url": "https://api.wisgate.ai/v1",
893      "api_key": "your_wisgate_api_key"
894    }
895  }
896}
897\`\`\`
898
899配置完成后,重启 Crush 应用,通过 \`/models\` 命令选择指定模型。
900
901## 三、使用场景
902
903- **终端内编码**:在项目目录中通过 TUI 直接处理代码生成和修改。
904- **文件操作**:让 Crush 读取项目上下文并辅助编辑文件。
905- **调试排错**:结合命令行输出、日志和测试结果生成修复建议。
906- **多任务处理**:把复杂开发需求拆分成多个终端内任务逐步完成。
907
908## 常见问题
909
910### \`/models\` 中没有看到配置的模型
911
912请确认配置文件路径正确、JSON 格式有效,并重启 Crush。
913
914### API 调用失败
915
916请确认 \`base_url\` 为 \`https://api.wisgate.ai/v1\`,\`api_key\` 已替换为 WisGate 控制台创建的 API Key。
917`,J=`# Cursor
918
919Connect WisGate models to Cursor through the OpenAI-compatible protocol for AI chat, code generation, debugging, refactoring, and project analysis.
920
921Due to Cursor's limitations, only users with a Cursor Pro subscription or above can configure custom models. Non-Pro users will encounter errors after configuration.
922
923## 1. Install Cursor
924
925Visit the Cursor website to download and install the version for your operating system.
926
927## 2. Create New Provider/Model
928
929In Cursor, open the "Models" section and click the "Add Custom Model" button.
930
931- Select **OpenAI protocol**
932- Configure **OpenAI API Key** (obtained from the WisGate dashboard)
933- In **Override OpenAI Base URL**, replace the default URL with: \`https://api.wisgate.ai/v1\`
934- Enter the model you want to use, such as \`GPT-5.5\`, \`GPT-4.7\`, or \`GPT-4.5\`
935- Note: In Cursor, you must use uppercase model names, e.g., \`GPT-5.5\`
936
937## 3. Save and Switch Model
938
939After configuration, save your settings and select the newly created Provider on the main page.
940
941## 4. Start Using
942
943With this setup, you can start using WisGate models for code generation, debugging, task analysis, and more.
944
945## 5. Use Cases
946
947- **Code completion and generation**: Generate code in Cursor based on the current file or project context.
948- **Error fixing**: Paste errors or select problematic code and ask the model to identify the cause.
949- **Refactoring suggestions**: Analyze functions, components, or modules and get refactoring guidance.
950- **Project Q&A**: Ask questions about the existing codebase to understand business logic and dependencies faster.
951
952## FAQ
953
954### The custom model is not available after saving
955
956Confirm that your Cursor plan supports custom models, and check that the API key, model name, and \`https://api.wisgate.ai/v1\` are correct.
957
958### The model name is not recognized
959
960Cursor can be sensitive to model name casing. Use the model name from the dashboard or model list, and use uppercase names when required.
961`,V=`# Cursor
962
963将 WisGate 模型通过 OpenAI 兼容协议接入 Cursor,用于 AI 对
963话、代码生成、调试、重构和项目分析。
964
965由于 Cursor 的限制,只有订阅了 Cursor 高级会员及以上的用户才支持自定义配置模型。若非高级会员,配置后会报错。
966
967## 1. 安装 Cursor
968
969访问 Cursor 官网下载并安装适合您的操作系统的版本。
970
971## 2. 创建新 Provider/Model
972
973在 Cursor 中,打开「Models」部分,并点击「Add Custom Model」按钮。
974
975- 选择 **OpenAI 协议**
976- 配置 **OpenAI API Key**(从 WisGate 控制台获取)
977- 在 **Override OpenAI Base URL** 中,将默认 URL 替换:\`https://api.wisgate.ai/v1\`
978- 输入您希望使用的模型,如 \`GPT-5.5\`、\`GPT-4.7\` 或 \`GPT-4.5\`
979- 注意:在 Cursor 中,需要输入模型的大写名称,如 \`GPT-5.5\`
980
981## 3. 保存并切换模型
982
983配置完成后,保存设置并在主页上选择您刚创建的 Provider。
984
985## 4. 开始使用
986
987通过该设置,您可以开始使用 WisGate 模型进行代码生成、调试、任务分析等工作。
988
989## 5. 使用场景
990
991- **代码补全与生成**:在 Cursor 中基于当前文件或项目上下文生成代码。
992- **错误修复**:粘贴报错或选中问题代码,让模型辅助定位原因。
993- **重构建议**:让模型分析函数、组件或模块结构,给出重构方案。
994- **项目问答**:围绕已有代码库提问,快速理解业务逻辑和依赖关系。
995
996## 常见问题
997
998### 自定义模型保存后不可用
999
1000请确认账号套餐支持自定义模型,并检查 API Key、模型名称和 \`https://api.wisgate.ai/v1\` 是否填写正确。
1001
1002### 模型名称无法识别
1003
1004Cursor 对模型名称大小写较敏感。请按照控制台或模型列表中的名称填写,必要时使用大写模型名。
1005`,Y=`# Factory Droid
1006
1007Factory Droid is an enterprise-grade AI coding agent that runs in your terminal, handling end-to-end software development workflows.
1008
1009Factory Droid requires successful account authentication before use.
1010
1011## Step 1: Install Factory Droid
1012
1013**macOS / Linux:**
1014
1015\`\`\`bash
1016curl -fsSL https://app.factory.ai/cli | sh
1017\`\`\`
1018
1019**Windows:**
1020
1021\`\`\`powershell
1022irm https://app.factory.ai/cli/windows | iex
1023\`\`\`
1024
1025## Step 2: Configure WisGate Models
1026
1027### 1. Get Your API Key
1028
1029Create and obtain your API Key from the WisGate dashboard.
1030
1031### 2. Configure Custom Model
1032
1033Configuration file location:
1034- macOS/Linux: \`~/.factory/settings.json\`
1035- Windows: \`%USERPROFILE%\\.factory\\settings.json\`
1036
1037#### Anthropic Protocol
1038
1039\`\`\`json
1040{
1041  "customModels": [
1042    {
1043      "displayName": "Claude [WisGate] - Sonnet",
1044      "model": "claude-sonnet-4-6",
1045      "baseUrl": "https://api.wisgate.ai",
1046      "apiKey": "your_api_key",
1047      "provider": "anthropic",
1048      "maxOutputTokens": 131072
1049    }
1050  ]
1051}
1052\`\`\`
1053
1054#### OpenAI Chat Completion Protocol
1055
1056\`\`\`json
1057{
1058  "customModels": [
1059    {
1060      "displayName": "GPT [WisGate] - OpenAI",
1061      "model": "gpt-5.5",
1062      "baseUrl": "https://api.wisgate.ai/v1",
1063      "apiKey": "your_api_key",
1064      "provider": "generic-chat-completion-api",
1065      "maxOutputTokens": 131072
1066    }
1067  ]
1068}
1069\`\`\`
1070
1071Replace \`your_api_key\` with your actual API Key. The API Key is stored locally only.
1072
1073For the Anthropic protocol, usually enter the Base URL \`https://api.wisgate.ai\`. If the tool asks for the full Messages API URL, use \`https://api.wisgate.ai/v1/messages\`.
1074
1075### 3. Start Factory Droid
1076
1077\`\`\`bash
1078cd /path/to/your/project
1079droid
1080\`\`\`
1081
1082On first launch, you will be guided through browser login.
1083
1084### 4. Select Your Model
1085
1086After droid starts, use the \`/model\` command to select the custom model you configured.
1087
1088## Step 3: Start Using
1089
1090After selecting the model, you can let Factory Droid handle end-to-end development tasks in the terminal, such as:
1091
1092- Analyze project structure and tasks
1093- Modify code and generate explanations
1094- Diagnose issues from failed tests
1095- Break down complex requirements and execute them step by step
1096
1097## FAQ
1098
1099### Should I use the Anthropic or OpenAI Chat Completion protocol?
1100
1101Use Base URL \`https://api.wisgate.ai\` for Claude-family or Anthropic Messages-compatible models. If the tool asks for the full request URL, use \`https://api.wisgate.ai/v1/messages\`. Use \`https://api.wisgate.ai/v1\` for GPT-family or OpenAI-compatible models.
1102
1103### The custom model does not appear
1104
1105Confirm that \`~/.factory/settings.json\` is in the correct path, the JSON format is valid, and restart \`droid\`.
1106`,Z=`# Factory Droid
1107
1108Factory Droid 是一款企业级 AI 编码代理,它运行在你的终端中,负责端到端的软件开发工作流。
1109
1110Factory Droid 需要成功登录 Factory Droid 账户认证后才能使用。
1111
1112## 步骤 1:安装 Factory Droid
1113
1114**macOS / Linux:**
1115
1116\`\`\`bash
1117curl -fsSL https://app.factory.ai/cli | sh
1118\`\`\`
1119
1120**Windows:**
1121
1122\`\`\`powershell
1123irm https://app.factory.ai/cli/windows | iex
1124\`\`\`
1125
1126## 步骤 2:配置 WisGate 模型
1127
1128### 1. 获取 API 密钥
1129
1130在 WisGate 控制台中创建并获取您的 API Key。
1131
1132### 2. 配置自定义模型
1133
1134配置文件位置:
1135- macOS/Linux:\`~/.factory/settings.json\`
1136- Windows:\`%USERPROFILE%\\.factory\\settings.json\`
1137
1138#### Anthropic 协议
1139
1140\`\`\`json
1141{
1142  "customModels": [
1143    {
1144      "displayName": "Claude [WisGate] - Sonnet",
1145      "model": "claude-sonnet-4-6",
1146      "baseUrl": "https://api.wisgate.ai",
1147      "apiKey": "your_api_key",
1148      "provider": "anthropic",
1149      "maxOutputTokens": 131072
1150    }
1151  ]
1152}
1153\`\`\`
1154
1155#### OpenAI Chat Completion 协议
1156
1157\`\`\`json
1158{
1159  "customModels": [
1160    {
1161      "displayName": "GPT [WisGate] - OpenAI",
1162      "model": "gpt-5.5",
1163      "baseUrl": "https://api.wisgate.ai/v1",
1164      "apiKey": "your_api_key",
1165      "provider": "generic-chat-completion-api",
1166      "maxOutputTokens": 131072
1167    }
1168  ]
1169}
1170\`\`\`
1171
1172将 \`your_api_key\` 替换为你的实际 API Key。API Key 仅保存在本地。
1173
1174Anthropic 协议通常填写 Base URL \`https://api.wisgate.ai\`。如果工具要求填写完整的 Messages API 地址,请使用 \`https://api.wisgate.ai/v1/messages\`。
1175
1176### 3. 启动 Factory Droid
1177
1178\`\`\`bash
1179cd /path/to/your/project
1180droid
1181\`\`\`
1182
1183首次启动时,系统会引导你通过浏览器登录。
1184
1185### 4. 选择你的模型
1186
1187droid 运行后,使用 \`/model\` 命令选择刚配置的自定义模型。
1188
1189## 步骤 3:开始使用
1190
1191选择模型后,可以让 Factory Droid 在终端中处理端到端开发任务,例如:
1192
1193- 分析项目结构和待办事项
1194- 修改代码并生成说明
1195- 根据测试失败信息定位问题
1196- 拆解复杂需求并逐步执行
1197
1198## 常见问题
1199
1200### 应该使用 Anthropic 还是 OpenAI Chat Completion 协议?
1201
1202Claude 系列或 Anthropic Messages 兼容模型使用 Base URL \`https://api.wisgate.ai\`;如果工具要求完整请求地址,则使用 \`https://api.wisgate.ai/v1/messages\`。GPT 系列或 OpenAI 兼容模型使用 \`https://api.wisgate.ai/v1\`。
1203
1204### 自定义模型没有出现
1205
1206请确认 \`~/.factory/settings.json\` 路径正确、JSON 格式有效,并重启 \`droid\`。
1207`,q=`# Goose
1208
1209Goose is an AI Agent tool that supports running locally or on the desktop, also available in CLI form. It supports multiple model integrations and uses the MCP protocol to connect external tools and APIs, automating engineering tasks such as code generation, debugging, testing, and deployment.
1210
1211## 1. Install Goose Desktop
1212
12131. Visit the Goose documentation page for quickstart
12142. Choose the appropriate installation method for your operating system
12153. Complete the installation
1216
1217## 2. Create New Provider
1218
12191. Open the Goose desktop application and enter the main interface
12202. Find and click "Create New Provider" in the left menu
12213. Follow the prompts to enter the required information
1222
1223## 3. Select Anthropic Protocol and Configure
1224
12251. During Provider creation, select the Anthropic protocol
12262. Fill in the following configuration:
1227   - **Base URL**: \`https://api.wisgate.ai\`
1228   - **API Key**: Enter your WisGate API Key
1229   - **Model**: Select GLM-5.1, GLM-5-Turbo, GLM-4.7, or GLM-4.5-air
12303. Save settings after configuration
1231
1232If Goose asks for the full Messages API URL instead of a Base URL, use \`https://api.wisgate.ai/v1/messages\`.
1233
1234## 4. Switch Model
1235
12361. Return to the Goose desktop main interface
12372. Find and click "Switch Models" at the bottom
12383. Select the newly created Provider from the dropdown list
12394. Ensure the new Provider is successfully switched as the active model
1240
1241## 5. Start Using
1242
1243After configuration and model switching, you can start using the Goose desktop with WisGate models. Enter your requests and Goose will automatically call the configured model to generate responses.
1244
1245## 6. Use Cases
1246
1247- **Engineering automation**: Use Goose with MCP tools for code generation, testing, and deployment-related tasks.
1248- **Local agent workflow**: Continue project conversations and execution in the desktop app or CLI.
1249- **Tool calling**: Connect external tools, APIs, or local services through MCP.
1250- **Debugging and verification**: Generate next-step suggestions from logs, test results, and file context.
1251
1252## FAQ
1253
1254### The Provider cannot call the model
1255
1256Confirm that the protocol is Anthropic, Base URL is \`https://api.wisgate.ai\`, and the API key is correct. If the interface asks for the full request URL, use \`https://api.wisgate.ai/v1/messages\`.
1257
1258### Goose still calls the old Provider after switching
1259
1260Return to "Switch Models" at the bottom of the main interface, select the new Provider again, and confirm that the current session switched successfully.
1261`,Q=`# Goose
1262
1263Goose 是一款 AI Agent 工具,支持在本地或桌面环境运行,也提供 CLI 形式。它支持多种模型接入,结合 MCP 协议连接外部工具与 API,自动化执行工程任务,如代码生成、调试、测试与部署。
1264
1265## 一、安装 Goose 桌面版
1266
12671. 访问 Goose 桌面版的官方文档页面,根据您的操作系统选择合适的安装方式
12682. 完成安装
1269
1270## 二、创建新 Provider
1271
12721. 打开 Goose 桌面版应用,进入主界面
12732. 找到并点击左侧菜单中的「创建新 Provider」
12743. 按照提示输入所需信息,完成新 Provider 的创建
1275
1276## 三、选择 Anthropic 协议并配置
1277
12781. 在创建 Provider 的过程中,选择 Anthropic 协议
12792. 填写以下必要的配置:
1280   - **Base URL**:\`https://api.wisgate.ai\`
1281   - **API Key**:输入您的 WisGate API 密钥
1282   - **Model**:选择 GLM-5.1、GLM-5-Turbo、GLM-4.7 或 GLM-4.5-air
12833. 配置完成后,保存设置
1284
1285如果 Goose 要求填写完整的 Messages API 地址,而不是 Base URL,请使用 \`https://api.wisgate.ai/v1/messages\`。
1286
1287## 四、切换模型
1288
12891. 回到 Goose 桌面版的主界面
12902. 在主界面最底部找到并点击「Switch Models」
12913. 在下拉列表中选择您刚才创建的新 Provider
12924. 确保新 Provider 已成功切换为当前使用模型
1293
1294## 五、开始使用
1295
1296配置完成并切换模型后,您就可以开始使用 Goose 桌面版与 WisGate 模型进行交互。输入您的请求,Goose 将根据您的配置自动调用模型并生成响应。
1297
1298## 六、使用场景
1299
1300- **工程自动化**:让 Goose 结合 MCP 工具执行代码生成、测试和部署相关任务。
1301- **本地 Agent 工作流**:在桌面端或 CLI 中围绕项目持续对话和执行。
1302- **工具调用**:通过 MCP 连接外部工具、API 或本地服务。
1303- **调试与验证**:根据日志、测试结果和文件上下文生成下一步操作建议。
1304
1305## 常见问题
1306
1307### Provider 创建后无法调用
1308
1309请确认协议选择为 Anthropic,Base URL 为 \`https://api.wisgate.ai\`,并检查 API Key 是否正确。如果界面要求完整请求地址,请填写 \`https://api.wisgate.ai/v1/messages\`。
1310
1311### 切换模型后仍调用旧 Provider
1312
1313请回到主界面底部的「Switch Models」重新选择新 Provider,并确认当前会话已经切换成功。
1314`,$=`# Hermes Agent
1315
1316Hermes Agent is a terminal-native autonomous coding agent built by Nous Research. It supports persistent memory, agent-created skills, messaging gateways, and multiple execution backends. You can connect Hermes Agent to WisGate through a custom OpenAI-compatible endpoint.
1317
1318## 1. Install Hermes Agent
1319
1320Follow the official Hermes Agent documentation to install it first. After installation, verify that the command is available:
1321
1322\`\`\`bash
1323hermes --version
1324\`\`\`
1325
1326## 2. Use the Interactive Model Selector
1327
1328Hermes recommends configuring models through the interactive selector:
1329
1330\`\`\`bash
1331hermes model
1332\`\`\`
1333
1334During the model setup flow, select a custom OpenAI-compatible endpoint and enter:
1335
1336- **Base URL**: \`https://api.wisgate.ai/v1\`
1337- **API Key**: Your WisGate API key
1338- **Model**: For example, \`gpt-5.5\`
1339
1340## 3. Manual Configuration
1341
1342Hermes usually stores secrets in \`~/.hermes/.env\`, while model and non-secret settings live in \`~/.hermes/config.yaml\`.
1343
1344Add this to \`~/.hermes/.env\`:
1345
1346\`\`\`env
1347OPENAI_API_KEY=your_wisgate_api_key
1348OPENAI_BASE_URL=https://api.wisgate.ai/v1
1349\`\`\`
1350
1351Configure the model in \`~/.hermes/config.yaml\`:
1352
1353\`\`\`yaml
1354model:
1355  provider: custom
1356  default: gpt-5.5
1357  base_url: https://api.wisgate.ai/v1
1358\`\`\`
1359
1360If your Hermes version uses a different field name than \`model.default\`, prefer the configuration generated by \`hermes model\`.
1361
1362## 4. Start Using
1363
1364\`\`\`bash
1365hermes
1366\`\`\`
1367
1368Or use the TUI:
1369
1370\`\`\`bash
1371hermes --tui
1372\`\`\`
1373
1374## Use Cases
1375
1376- **Autonomous development tasks**: Let Hermes run multi-step development tasks locally, in Docker, over SSH, or on remote backends.
1377- **Persistent memory**: Preserve project preferences, experience, and skills across sessions.
1378- **Messaging gateway**: Trigger agent workflows from Telegram, Discord, Slack, and other platforms.
1379- **Auxiliary models**: Configure cheaper WisGate models for compression, vision, titles, and other lightweight tasks.
1380
1381## FAQ
1382
1383### Hermes cannot find the API key
1384
1385Confirm that \`OPENAI_API_KEY\` exists in \`~/.hermes/.env\`, then restart Hermes.
1386
1387### Model or context errors occur
1388
1389Confirm that the model ID is supported by WisGate, and prefer a model with a large enough context window. The OpenAI-compatible Base URL is \`https://api.wisgate.ai/v1\`.
1390`,X=`# Hermes Agent
1391
1392Hermes Agent 是 Nous Research 构建的终端原生自治编程代理,支持持久记忆、技能创建、消息网关和多种运行后端。通过自定义 OpenAI 兼容端点,可以将 Hermes Agent 接入 WisGate。
1393
1394## 一、安装 Hermes Agent
1395
1396请先按照 Hermes Agent 官方文档完成安装。安装完成后确认命令可用:
1397
1398\`\`\`bash
1399hermes --version
1400\`\`\`
1401
1402## 二、使用交互式模型选择器
1403
1404Hermes 推荐使用交互式模型选择器配置模型:
1405
1406\`\`\`bash
1407hermes model
1408\`\`\`
1409
1410在模型配置流程中选择自定义 OpenAI-compatible endpoint,并填写:
1411
1412- **Base URL**:\`https://api.wisgate.ai/v1\`
1413- **API Key**:您的 WisGate API Key
1414- **Model**:例如 \`gpt-5.5\`
1415
1416## 三、手动配置
1417
1418Hermes 的密钥通常放在 \`~/.hermes/.env\`,模型与非密钥配置放在 \`~/.hermes/config.yaml\`。
1419
1420在 \`~/.hermes/.env\` 中写入:
1421
1422\`\`\`env
1423OPENAI_API_KEY=your_wisgate_api_key
1424OPENAI_BASE_URL=https://api.wisgate.ai/v1
1425\`\`\`
1426
1427在 \`~/.hermes/config.yaml\` 中配置模型:
1428
1429\`\`\`yaml
1430model:
1431  provider: custom
1432  default: gpt-5.5
1433  base_url: https://api.wisgate.ai/v1
1434\`\`\`
1435
1436如果您的 Hermes 版本使用 \`model.default\` 以外的字段名,请优先以 \`hermes model\` 生成的配置为准。
1437
1438## 四、开始使用
1439
1440\`\`\`bash
1441hermes
1442\`\`\`
1443
1444或使用 TUI:
1445
1446\`\`\`bash
1447hermes --tui
1448\`\`\`
1449
1450## 使用场景
1451
1452- **自治开发任务**:让 Hermes 在本地、Docker、SSH 或远程后端执行多步骤开发任务。
1453- **持久记忆**:跨会话保留项目偏好、经验和技能。
1454- **消息网关**:通过 Telegram、Discord、Slack 等平台触发 Agent 工作流。
1455- **辅助模型**:为压缩、视觉、标题等轻量任务单独配置更便宜的 WisGate 模型。
1456
1457## 常见问题
1458
1459### 找不到 API Key
1460
1461请确认 \`~/.hermes/.env\` 中有 \`OPENAI_API_KEY\`,并重启 Hermes。
1462
1463### 模型或上下文报错
1464
1465请确认模型 ID 为 WisGate 支持的模型,并优先选择上下文窗口足够大的模型。OpenAI 兼容 Base URL 使用 \`https://api.wisgate.ai/v1\`。
1466`,nn=`# Junie CLI
1467
1468Junie CLI is JetBrains' agentic coding command-line tool. It supports BYOK and custom model profiles. With a custom LLM profile, Junie CLI can connect to WisGate through the OpenAI-compatible interface.
1469
1470## 1. Install Junie CLI
1471
1472On Linux / macOS, use the install script:
1473
1474\`\`\`bash
1475curl -fsSL https://junie.jetbrains.com/install.sh | bash
1476\`\`\`
1477
1478Verify the installation:
1479
1480\`\`\`bash
1481junie --version
1482\`\`\`
1483
1484## 2. Create a WisGate Custom Model Profile
1485
1486Junie CLI discovers custom models from \`$JUNIE_HOME/models/*.json\` or project-level \`.junie/models/*.json\`. If \`JUNIE_HOME\` is not set, use the default user directory \`~/.junie\`:
1487
1488\`\`\`bash
1489mkdir -p ~/.junie/models
1490\`\`\`
1491
1492Create \`~/.junie/models/wisgate.json\`:
1493
1494\`\`\`json
1495{
1496  "id": "gpt-5.5",
1497  "baseUrl": "https://api.wisgate.ai/v1/chat/completions",
1498  "apiType": "OpenAICompletion",
1499  "apiKey": "your_wisgate_api_key",
1500  "temperature": 0.7,
1501  "fasterModel": {
1502    "id": "gpt-4.7"
1503  }
1504}
1505\`\`\`
1506
1507If you want to share the model profile inside a project, place it at \`.junie/models/wisgate.json\`, but do not commit a real API key.
1508
1509## 3. Start Junie
1510
1511Navigate to your project directory and select the model:
1512
1513\`\`\`bash
1514cd /path/to/your/project
1515junie --model wisgate
1516\`\`\`
1517
1518You can also start Junie first, then use \`/model\` to view and switch models.
1519
1520## Use Cases
1521
1522- **Interactive coding**: Review, write, and modify code from the terminal.
1523- **File context**: Use \`@\` to attach files or folders to the request context.
1524- **Automated review**: Run headless code quality checks and fix suggestions.
1525- **Team profiles**: Share a \`.junie/models/wisgate.json\` template without secrets.
1526
1527## FAQ
1528
1529### Junie cannot find the custom model
1530
1531Confirm that the JSON file is in \`~/.junie/models/\`, \`$JUNIE_HOME/models/\`, or \`.junie/models/\`, and that the filename matches the \`--model\` value.
1532
1533### Requests fail
1534
1535Confirm that \`baseUrl\` is the full Chat Completions URL \`https://api.wisgate.ai/v1/chat/completions\`, \`apiType\` is \`OpenAICompletion\`, and the API key is correct.
1536`,en=`# Junie CLI
1537
1538Junie CLI 是 JetBrains 提供的 agentic 编程命令行工具,支持 BYOK 和自定义模型 Profile。通过自定义 LLM Profile,可以把 Junie CLI 接入 WisGate 的 OpenAI 兼容接口。
1539
1540## 一、安装 Junie CLI
1541
1542Linux / macOS 可使用安装脚本:
1543
1544\`\`\`bash
1545curl -fsSL https://junie.jetbrains.com/install.sh | bash
1546\`\`\`
1547
1548验证安装:
1549
1550\`\`\`bash
1551junie --version
1552\`\`\`
1553
1554## 二、创建 WisGate 自定义模型 Profile
1555
1556Junie CLI 会从 \`$JUNIE_HOME/models/*.json\` 或项目内 \`.junie/models/*.json\` 读取自定义模型。若未设置 \`JUNIE_HOME\`,可使用默认用户目录 \`~/.junie\`:
1557
1558\`\`\`bash
1559mkdir -p ~/.junie/models
1560\`\`\`
1561
1562新增 \`~/.junie/models/wisgate.json\`:
1563
1564\`\`\`json
1565{
1566  "id": "gpt-5.5",
1567  "baseUrl": "https://api.wisgate.ai/v1/chat/completions",
1568  "apiType": "OpenAICompletion",
1569  "apiKey": "your_wisgate_api_key",
1570  "temperature": 0.7,
1571  "fasterModel": {
1572    "id": "gpt-4.7"
1573  }
1574}
1575\`\`\`
1576
1577如果需要在项目中共享模型配置,请把 profile 放到 \`.junie/models/wisgate.json\`,但不要把真实 API Key 提交到代码仓库。
1578
1579## 三、启动 Junie
1580
1581进入项目目录并指定模型:
1582
1583\`\`\`bash
1584cd /path/to/your/project
1585junie --model wisgate
1586\`\`\`
1587
1588也可以先启动 Junie,在会话中使用 \`/model\` 查看并切换模型。
1589
1590## 使用场景
1591
1592- **交互式编码**:在终端中让 Junie 审查、编写和修改代码。
1593- **文件上下文**:使用 \`@\` 附加文件或目录,让模型读取指定上下文。
1594- **自动化审查**:在 headless 模式中执行代码质量检查和修复建议。
1595- **团队 Profile**:把不含密钥的 \`.junie/models/wisgate.json\` 作为团队模板。
1596
1597## 常见问题
1598
1599### Junie 没有发现自定义模型
1600
1601请确认 JSON 文件位于 \`~/.junie/models/\`、\`$JUNIE_HOME/models/\` 或 \`.junie/models/\`,并确认文件名和 \`--model\` 参数一致。
1602
1603### 请求失败
1604
1605请确认 \`baseUrl\` 使用完整 Chat Completions 地址 \`https://api.wisgate.ai/v1/chat/completions\`,\`apiType\` 为 \`OpenAICompletion\`,API Key 正确。
1606`,tn=`# Kilo Code
1607
1608Kilo Code is a powerful VS Code extension that supports MCP (Model Context Protocol), helping you with code generation, debugging, and project management in the editor for more efficient intelligent development.
1609
1610When using WisGate, configure the API endpoint \`https://api.wisgate.ai/v1\`.
1611
1612## 1. Install Kilo Code Extension
1613
1614### 1. Open Extension Marketplace
1615
16161. Open VS Code
16172. Click the extension marketplace icon on the left
16183. Search for \`Kilo Code\` in the search box
16194. Find the \`Kilo Code\` extension
1620
1621### 2. Install Extension
1622
16231. Click the \`Install\` button
16242. After installation, select trust developer
1625
1626## 2. Configure API Settings
1627
1628### 1. Select API Key Method
1629
1630Select \`Use your own API key\`
1631
1632### 2. Enter Configuration
1633
1634- **API Provider**: Select \`Z AI\`
1635- **Z AI Entrypoint**: Select WisGate endpoint (\`https://api.wisgate.ai/v1\`)
1636- **Z AI API Key**: Enter your WisGate API Key
1637- **Model**: Select \`gpt-5.5\` or any model from the list
1638
1639## 3. Start Using
1640
1641After configuration, you can enter your requirements in the input box:
1642
1643- Analyze data table structures
1644- Calculate statistics and averages
1645- Generate and optimize SQL queries
1646- Code generation and refactoring
1647- Project analysis and documentation
1648
1649## 4. Usage Tips
1650
1651- Kilo Code is suitable for code generation, SQL work, documentation, and project analysis inside VS Code.
1652- If the task needs MCP tools, confirm that the related MCP services are enabled in Kilo Code.
1653- If the target model is not in the list, check whether it is available in the WisGate dashboard.
1654
1655## FAQ
1656
1657### Connection fails
1658
1659Confirm that \`Z AI\` is selected, Entrypoint is \`https://api.wisgate.ai/v1\`, and the API key is correct.
1660
1661### The model is slow or interrupted
1662
1663Try another model, or split complex tasks into separate analysis, modification, and verification stages.
1664`,on=`# Kilo Code
1665
1666Kilo Code 是一个功能强大的 VS Code 插件,支持 MCP(Model Context Protocol),能够帮助你在编辑器中进行代码生成、调试和项目管理等任务,实现更高效的智能开发。
1667
1668使用 WisGate 时,配置 API 端点 \`https://api.wisgate.ai/v1\` 即可。
1669
1670## 一、安装 Kilo Code 插件
1671
1672### 1. 打开插件市场
1673
16741. 打开 VS Code
16752. 点击左侧插件市场图标
16763. 在搜索框中输入 \`Kilo Code\`
16774. 找到 \`Kilo Code\` 插件
1678
1679### 2. 安装插件
1680
16811. 点击 \`Install\` 按钮进行安装
16822. 安装完成后,选择信任开发者
1683
1684## 二、配置 API 设置
1685
1686### 1. 选择 API Key 方式
1687
1688选择 \`使用你自己的 API 秘钥\`
1689
1690### 2. 填入配置信息
1691
1692- **API Provider**:选择 \`Z AI\`
1693- **Z AI Entrypoint**:选择 WisGate 端点 (\`https://api.wisgate.ai/v1\`)
1694- **Z AI API Key**:填入您的 WisGate API Key
1695- **Model**:选择 \`gpt-5.5\` 或者列表中您想使用的模型
1696
1697## 三、开始使用
1698
1699配置完成后,您可以在输入框中输入需求,让 AI 模型帮助您完成各种任务:
1700
1701- 分析数据表结构
1702- 统计数据和计算平均值
1703- 生成和优化 SQL 查询
1704- 代码生成和重构
1705- 项目分析和文档编写
1706
1707## 四、使用建议
1708
1709- 适合在 VS Code 内处理代码生成、SQL、文档和项目分析任务。
1710- 如果任务需要调用 MCP 工具,请先确认 Kilo Code 中相关 MCP 服务已启用。
1711- 模型列表中没有目标模型时,可先确认 WisGate 控制台中该模型是否可用。
1712
1713## 常见问题
1714
1715### 连接失败
1716
1717请确认选择的是 \`Z AI\`,Entrypoint 为 \`https://api.wisgate.ai/v1\`,并检查 API Key 是否正确。
1718
1719### 模型返回慢或中断
1720
1721可尝试更换模型,或把复杂任务拆成“分析、修改、验证”几个阶段分别执行。
1722`,an=`# Lingma
1723
1724Lingma IDE is an Agentic programming platform for enterprises and developers. Lingma also provides a JetBrains plugin. After connecting WisGate, you can use a unified model service for code generation, project Q&A, debugging, and refactoring.
1725
1726Due to Lingma's limitations, only users with a Lingma Personal Pro subscription can configure custom models.
1727
1728## 1. Install and Use
1729
17301. Visit the Lingma website to download and install Lingma IDE (v0.6.0 or later)
17312. Complete the initial setup after first launch
17323. Log in to see your account status in Lingma IDE
1733
1734## 2. Connect WisGate
1735
17361. Open Lingma settings in the top right corner, select **Models**, and click **Add**
17372. Model configuration:
1738   - Provider: Select the WisGate provider
1739   - Type: Select WisGate
1740   - Model: Select a model from the dropdown
1741   - API Key: Enter your WisGate API Key
17423. Click **Add** to complete model configuration after validation
17434. Select the model in the Lingma chat to start using it
1744
1745Note: The Lingma JetBrains plugin also supports WisGate.
1746
1747## 3. Use Cases
1748
1749- **Enterprise project Q&A**: Ask questions about the current repository, business modules, and API logic.
1750- **Code generation**: Generate functions, components, APIs, and automation scripts.
1751- **Code review**: Check potential errors, edge cases, and maintainability issues.
1752- **Plugin workflow**: Reuse the same WisGate model configuration inside JetBrains IDEs.
1753
1754## FAQ
1755
1756### Why can't I find the model option in Lingma settings?
1757
1758You need to log in first before you can chat and configure models.
1759
1760### Model validation fails
1761
1762Confirm that your account supports custom models, the API key is correct, and you selected the built-in WisGate provider and model type when available.
1763`,sn=`# 灵码(Lingma)
1764
1765灵码 IDE 是面向企业及开发者的 Agentic 编程平台,同时提供 JetBrains 插件能力。接入 WisGate 后,可以在灵码中使用统一模型完成代码生成、项目问答、调试和重构。
1766
1767由于灵码的限制,只有订阅了灵码个人专业版的用户才能支持自定义配置模型。
1768
1769## 1. 安装与使用
1770
17711. 前往灵码官网下载并安装灵码 IDE(v0.6.0 版本及以上)
17722. 初次启动灵码后,需要完成初始配置
17733. 登录后即可在灵码 IDE 中看到已登录状态
1774
1775## 2. 接入 WisGate
1776
17771. 在界面右上角打开灵码设置,选择**模型**,单击**添加**
17782. 模型配置信息:
1779   - 提供商:选择 WisGate 对应的提供商
1780   - 类型:选择 WisGate
1781   - 模型:在下拉菜单中选择模型
1782   - API Key:填写 WisGate API Key
17833. 单击**添加**,通过校验后即可完成模型配置
17844. 在灵码对话框中,选择相应的模型即可开始使用
1785
1786备注:灵码 JetBrains 插件也支持 WisGate。
1787
1788## 3. 使用场景
1789
1790- **企业项目问答**:围绕当前仓库、业务模块和接口逻辑提问。
1791- **代码生成**:生成函数、组件、接口和自动化脚本。
1792- **代码审查**:检查潜在错误、边界情况和可维护性问题。
1793- **插件开发流**:在 JetBrains IDE 中复用同一套 WisGate 模型配置。
1794
1795## 常见问题
1796
1797### 为什么在灵码设置中找不到模型选项?
1798
1799您需要先完成登录,才能够进行对话和配置模型。
1800
1801### 添加模型校验失败怎么办?
1802
1803请确认当前账号支持自定义模型,API Key 填写正确,并优先选择灵码内置的 WisGate 提供商和模型类型。
1804`,rn=`# OpenClaw
1805
1806OpenClaw is an open-source AI agent platform for bringing conversational agents to Telegram, Discord, Slack, Signal, iMessage, WhatsApp, and other messaging channels. WisGate can be connected as an OpenAI-compatible model service.
1807
1808## 1. Install OpenClaw
1809
1810Follow the official OpenClaw documentation to install it first. After installation, verify that the command is available:
1811
1812\`\`\`bash
1813openclaw --version
1814\`\`\`
1815
1816## 2. Use the Setup Wizard
1817
1818If your OpenClaw version supports a custom OpenAI-compatible provider, run:
1819
1820\`\`\`bash
1821openclaw onboard
1822\`\`\`
1823
1824During setup, choose a custom or OpenAI-compatible provider and enter:
1825
1826- **Base URL**: \`https://api.wisgate.ai/v1\`
1827- **API Key**: Your WisGate API key
1828- **Model**: For example, \`gpt-5.5\`
1829
1830## 3. Manual Configuration
1831
1832Some OpenClaw versions support OpenAI-compatible endpoints through \`config.yaml\`. The config file may be located at \`~/.openclaw/config.yaml\` or under the installation directory.
1833
1834\`\`\`yaml
1835llm:
1836  provider: openai
1837  api_key: your_wisgate_api_key
1838  base_url: https://api.wisgate.ai/v1
1839  model: gpt-5.5
1840\`\`\`
1841
1842Save the file and restart OpenClaw:
1843
1844\`\`\`bash
1845openclaw gateway run
1846\`\`\`
1847
1848If your version only supports the built-in openai provider and does not support a custom Base URL, it cannot connect to WisGate directly. Upgrade to a version that supports custom OpenAI-compatible endpoints.
1849
1850## Use Cases
1851
1852- **Multi-channel agents**: Use WisGate models across Telegram, Discord, Slack, and other channels.
1853- **Team assistants**: Configure different models or workflows for different channels.
1854- **Automation tasks**: Trigger code analysis, summaries, and tool calls through messages.
1855- **Cost control**: Use lighter models for simple messages and stronger models for complex tasks.
1856
1857## FAQ
1858
1859### There is no custom provider option
1860
1861Check your OpenClaw version. Older versions may only support the built-in openai provider and cannot directly accept a WisGate Base URL.
1862
1863### The API key does not work
1864
1865Confirm that the config file path is correct, \`base_url\` is \`https://api.wisgate.ai/v1\`, and restart the gateway.
1866`,ln=`# OpenClaw
1867
1868OpenClaw 是一个开源 AI Agent 平台,可把对话式 Agent 接入 Telegram、Discord、Slack、Signal、iMessage、WhatsApp 等多个消息渠道。WisGate 可作为 OpenAI 兼容模型服务接入 OpenClaw。
1869
1870## 一、安装 OpenClaw
1871
1872请先按照 OpenClaw 官方文档完成安装。安装后确认命令可用:
1873
1874\`\`\`bash
1875openclaw --version
1876\`\`\`
1877
1878## 二、使用配置向导
1879
1880如果您的 OpenClaw 版本支持自定义 OpenAI-compatible provider,可以运行:
1881
1882\`\`\`bash
1883openclaw onboard
1884\`\`\`
1885
1886在配置过程中选择自定义或 OpenAI 兼容 Provider,并填写:
1887
1888- **Base URL**:\`https://api.wisgate.ai/v1\`
1889- **API Key**:您的 WisGate API Key
1890- **Model**:例如 \`gpt-5.5\`
1891
1892## 三、手动配置
1893
1894部分 OpenClaw 版本支持通过 \`config.yaml\` 配置 OpenAI 兼容端点。配置文件位置可能为 \`~/.openclaw/config.yaml\` 或安装目录下的 \`config.yaml\`。
1895
1896\`\`\`yaml
1897llm:
1898  provider: openai
1899  api_key: your_wisgate_api_key
1900  base_url: https://api.wisgate.ai/v1
1901  model: gpt-5.5
1902\`\`\`
1903
1904保存后重启 OpenClaw:
1905
1906\`\`\`bash
1907openclaw gateway run
1908\`\`\`
1909
1910如果您的版本只支持内置 openai provider,而不支持自定义 Base URL,则不能直接接入 WisGate;请升级到支持自定义 OpenAI-compatible endpoint 的版本。
1911
1912## 使用场景
1913
1914- **多渠道 Agent**:在 Telegram、Discord、Slack 等渠道中统一调用 WisGate 模型。
1915- **团队助理**:为不同频道配置不同模型或不同工作流。
1916- **自动化任务**:让 Agent 在消息中触发代码分析、摘要和工具调用。
1917- **成本控制**:用较轻模型处理简单消息,用更强模型处理复杂任务。
1918
1919## 常见问题
1920
1921### 没有自定义 Provider 选项
1922
1923请检查 OpenClaw 版本。旧版本可能只支持内置 openai provider,无法直接填写 WisGate Base URL。
1924
1925### API Key 不生效
1926
1927确认配置文件位置正确,\`base_url\` 为 \`https://api.wisgate.ai/v1\`,并重启 gateway。
1928`,dn=`# OpenCode
1929
1930OpenCode is a CLI + TUI AI programming agent that runs in the terminal and also provides IDE plugin integration. By connecting OpenCode to WisGate, developers can use WisGate's OpenAI-compatible endpoint for code generation, debugging, project analysis, file operations, and cross-project collaboration.
1931
1932When using WisGate, configure the API endpoint \`https://api.wisgate.ai/v1\`.
1933
1934## 1. Install OpenCode
1935
1936The easiest way to install OpenCode is using the official install script:
1937
1938\`\`\`bash
1939curl -fsSL https://opencode.ai/install | bash
1940\`\`\`
1941
1942You can also install via npm:
1943
1944\`\`\`bash
1945npm install -g opencode-ai
1946\`\`\`
1947
1948## 2. Configure WisGate Models
1949
1950### 1. Get Your API Key
1951
1952Create and obtain your API Key from the WisGate dashboard.
1953
1954### 2. Run \`opencode auth login\`
1955
1956\`\`\`bash
1957$ opencode auth login
1958\`\`\`
1959
1960Select the WisGate provider option.
1961
1962### 3. Enter Your API Key
1963
1964Enter your WisGate API Key when prompted.
1965
1966### 4. Start OpenCode
1967
1968\`\`\`bash
1969$ opencode
1970\`\`\`
1971
1972After entering OpenCode, use the \`/models\` command to select the model you need. If you need to reconnect a provider, use \`/connect\` to check or switch the current configuration.
1973
1974## WisGate Use Cases
1975
1976- **Terminal coding**: Run OpenCode inside a project directory to read files, generate code, and modify the codebase.
1977- **Project analysis**: Analyze folder structure, dependencies, error logs, and refactoring risks.
1978- **Multi-model switching**: Use \`/models\` in OpenCode to choose a model that fits the task.
1979- **Team configuration**: Standardize on \`https://api.wisgate.ai/v1\` and one API key management flow to reduce switching costs across model providers.
1980
1981### Legacy Version Configuration
1982
1983If your OpenCode version doesn't have the WisGate provider option, configure it in \`~/.config/opencode/opencode.json\`:
1984
1985\`\`\`json
1986{
1987  "$schema": "https://opencode.ai/config.json",
1988  "provider": {
1989    "wisgate": {
1990      "npm": "@ai-sdk/openai-compatible",
1991      "name": "WisGate",
1992      "options": {
1993        "baseURL": "https://api.wisgate.ai/v1"
1994      },
1995      "models": {
1996        "claude-sonnet-4-6": {
1997          "name": "Claude Sonnet"
1998        }
1999      }
2000    }
2001  }
2002}
2003\`\`\`
2004
2005Save the configuration and run:
2006
2007\`\`\`bash
2008$ opencode
2009\`\`\`
2010
2011If the connection fails, verify that your API key is correct, confirm that \`baseURL\` is \`https://api.wisgate.ai/v1\`, and use \`/connect\` to select the current provider again.
2012`,cn=`# OpenCode
2013
2014OpenCode 既是一款在终端中运行的 CLI + TUI AI 编程代理工具,也提供 IDE 的插件集成。通过接入 WisGate,OpenCode 可以使用 OpenAI 兼容接口完成代码生成、调试、项目分析、文件操作与跨项目协作等任务。
2015
2016使用 WisGate 时,配置 API 端点 \`https://api.wisgate.ai/v1\` 即可。
2017
2018## 一、安装 OpenCode
2019
2020安装 OpenCode 最简单的方式是使用官方安装脚本:
2021
2022\`\`\`bash
2023curl -fsSL https://opencode.ai/install | bash
2024\`\`\`
2025
2026你也可以使用 npm 安装:
2027
2028\`\`\`bash
2029npm install -g opencode-ai
2030\`\`\`
2031
2032## 二、配置 WisGate 模型
2033
2034### 1. 获取 API 密钥
2035
2036在 WisGate 控制台中创建并获取您的 API Key。
2037
2038### 2. 运行 \`opencode auth login\` 并选择 WisGate
2039
2040\`\`\`bash
2041$ opencode auth login
2042\`\`\`
2043
2044在 Select provider 中选择 WisGate 对应的选项。
2045
2046### 3. 输入您的 API Key
2047
2048在提示中输入您的 WisGate API Key。
2049
2050### 4. 运行 \`opencode\` 启动 OpenCode
2051
2052\`\`\`bash
2053$ opencode
2054\`\`\`
2055
2056进入 OpenCode 后,使用 \`/models\` 命令选择所需模型。如果需要重新连接提供商,可使用 \`/connect\` 检查或切换当前配置。
2057
2058## WisGate 使用场景
2059
2060- **终端内编码**:直接在项目目录中让 OpenCode 读取文件、生成代码并修改项目。
2061- **项目分析**:分析目录结构、依赖关系、错误日志和重构风险。
2062- **多模型切换**:在 OpenCode 中使用 \`/models\` 按任务选择更适合的模型。
2063- **团队统一配置**:团队统一使用 \`https://api.wisgate.ai/v1\` 和同一套 API Key 管理方式,减少不同模型服务商之间的切换成本。
2064
2065### 低版本配置
2066
2067若您使用的低版本 OpenCode 内无 WisGate Provider 选项,建议升级版本。或者在 \`~/.config/opencode/opencode.json\` 中配置:
2068
2069\`\`\`json
2070{
2071  "$schema": "https://opencode.ai/config.json",
2072  "provider": {
2073    "wisgate": {
2074      "npm": "@ai-sdk/openai-compatible",
2075      "name": "WisGate",
2076      "options": {
2077        "baseURL": "https://api.wisgate.ai/v1"
2078      },
2079      "models": {
2080        "claude-sonnet-4-6": {
2081          "name": "Claude Sonnet"
2082        }
2083      }
2084    }
2085  }
2086}
2087\`\`\`
2088
2089保存配置后重新运行:
2090
2091\`\`\`bash
2092$ opencode
2093\`\`\`
2094
2095如果连接失败,请确认 API Key 是否正确、\`baseURL\` 是否为 \`https://api.wisgate.ai/v1\`,并通过 \`/connect\` 重新选择当前提供商。
2096`,pn=`# Roo Code
2097
2098Roo Code is an intelligent VS Code extension that helps you with project analysis, code generation, and refactoring, making the development process easier and more efficient.
2099
2100When using WisGate, configure the API endpoint \`https://api.wisgate.ai/v1\`.
2101
2102## 1. Install Roo Code Extension
2103
2104### 1. Open Extension Marketplace
2105
21061. Open VS Code
21072. Click the extension marketplace icon on the left
21083. Search for \`Roo Code\` in the search box
21094. Find the \`Roo Code\` extension
2110
2111### 2. Install Extension
2112
21131. Click the \`Install\` button
21142. After installation, select trust developer
2115
2116## 2. Configure API Settings
2117
2118- **API Provider**: Select \`Z AI\`
2119- **Z AI Entrypoint**: Select WisGate endpoint (\`https://api.wisgate.ai/v1\`)
2120- **Z AI API Key**: Enter your WisGate API Key
2121- **Model**: Select \`gpt-5.5\` or any model from the list
2122
2123## 3. Permissions and Usage
2124
2125### 1. Configure Permissions
2126
2127Select the permissions you want to allow:
2128- File read/write operations
2129- Auto-approve execution
2130- Project access permissions
2131
2132### 2. Start Using
2133
2134Enter your requirements in the input box. Roo Code can help you with:
2135- Summarizing current project structure
2136- Analyzing key modules and features
2137- Code refactoring and optimization
2138- Generating documentation and comments
2139- Issue diagnosis and fix suggestions
2140
2141## 4. Usage Tips
2142
2143- Be careful with auto-approve execution. Start with file read/write permissions and expand gradually.
2144- Before asking Roo Code to modify a project, specify target files, directories that must not be changed, and acceptance criteria.
2145- For cross-file refactoring, ask it to output a plan before making changes.
2146
2147## FAQ
2148
2149### The model cannot be called
2150
2151Check that the \`Z AI\` provider, \`https://api.wisgate.ai/v1\`, and your WisGate API key are configured correctly.
2152
2153### There are many permission prompts
2154
2155This is Roo Code's safety mechanism. You can gradually enable auto-approval based on task risk, but avoid allowing unknown commands directly.
2156`,un=`# Roo Code
2157
2158Roo Code 是一个智能的 VS Code 插件,能够帮助你完成项目分析、代码生成和重构等任务,让开发过程更加轻松高效。
2159
2160使用 WisGate 时,配置 API 端点 \`https://api.wisgate.ai/v1\` 即可。
2161
2162## 一、安装 Roo Code 插件
2163
2164### 1. 打开插件市场
2165
21661. 打开 VS Code
21672. 点击左侧插件市场图标
21683. 在搜索框中输入 \`Roo Code\`
21694. 找到 \`Roo Code\` 插件
2170
2171### 2. 安装插件
2172
21731. 点击 \`Install\` 按钮进行安装
21742. 安装完成后,选择信任开发者
2175
2176## 二、配置 API 设置
2177
2178请按照以下配置填入相关信息:
2179
2180- **API Provider**:选择 \`Z AI\`
2181- **Z AI Entrypoint**:选择 WisGate 端点 (\`https://api.wisgate.ai/v1\`)
2182- **Z AI API Key**:填入您的 WisGate API Key
2183- **Model**:选择 \`gpt-5.5\` 或者列表中您想使用的模型
2184
2185## 三、权限设置和使用
2186
2187### 1. 配置权限
2188
2189根据您的具体需求,选择允许的权限:
2190- 文件读写操作
2191- 自动批准执行
2192- 项目访问权限
2193
2194### 2. 开始使用
2195
2196在输入框中输入您的需求,Roo Code 可以帮助您完成:
2197- 总结当前项目结构
2198- 分析重点模块和功能
2199- 代码重构和优化
2200- 生成文档和注释
2201- 问题诊断和修复建议
2202
2203## 四、使用建议
2204
2205- 对自动批准执行保持谨慎,建议先从文件读写权限开始逐步放开。
2206- 让 Roo Code 修改项目前,先说明目标文件、禁止修改的目录和验收标准。
2207- 对跨文件重构任务,建议要求它先输出计划,再执行修改。
2208
2209## 常见问题
2210
2211### 无法调用模型
2212
2213请检查 \`Z AI\` Provider、\`https://api.wisgate.ai/v1\` 和 WisGate API Key 是否填写正确。
2214
2215### 权限弹窗较多
2216
2217这是 Roo Code 的安全机制。可根据任务风险逐步开启自动批准,但不要对未知命令直接放开执行权限。
2218`,hn=`# TRAE
2219
2220TRAE is an intelligent IDE for developers, integrating AI Q&A, inline code completion, and Agentic Coding Workflow. After connecting WisGate, you can use models in TRAE for code generation, debugging, and task analysis.
2221
2222## 1. Install TRAE
2223
2224Visit the TRAE website to download and install the version for your operating system.
2225
2226## 2. Configure WisGate
2227
2228### 1. Add Custom Model
2229
2230In TRAE, click the **model switcher dropdown** or go to **Settings/Models**, and select add model.
2231
2232### 2. Select Model Provider
2233
2234Select the WisGate model provider.
2235
2236### 3. Select Model
2237
2238Choose or configure the model you want to use.
2239
2240### 4. Enter Your API Key
2241
2242Add your WisGate API Key.
2243
2244If the interface asks for an endpoint URL, use the OpenAI-compatible endpoint:
2245
2246\`\`\`text
2247https://api.wisgate.ai/v1
2248\`\`\`
2249
2250## 3. Save and Switch Models
2251
2252After configuration, save settings and select your newly created model in custom models. Done! You can now use WisGate models in TRAE for code generation, debugging, task analysis, and more.
2253
2254## 4. Use Cases
2255
2256- **AI Q&A**: Ask technical questions about the current project, file, or error.
2257- **Code generation**: Generate functions, APIs, pages, and scripts from requirements.
2258- **Debugging**: Analyze exceptions and context, then get fix suggestions.
2259- **Task planning**: Break complex development requests into actionable steps.
2260
2261## FAQ
2262
2263### The model configuration cannot be saved
2264
2265Check that your API key is valid and confirm that the endpoint is \`https://api.wisgate.ai/v1\`.
2266
2267### The output does not match expectations
2268
2269Make sure the selected model is the WisGate model you configured, and include project context, target files, and expected output format in your prompt.
2270`,gn=`# TRAE
2271
2272TRAE 是一款面向开发者打造的智能 IDE,融合 AI 问答、行内代码补全及 Agentic Coding Workflow 等功能。接入 WisGate 后,可以在 TRAE 中统一使用模型进行代码生成、调试和任务分析。
2273
2274## 一、安装 TRAE
2275
2276访问 TRAE 官网下载并安装适合您的操作系统的版本。
2277
2278## 二、配置 WisGate
2279
2280### 1. 添加自定义模型
2281
2282在 TRAE 中,点击**模型切换下拉菜单**或**进入 Settings/Models**,选择添加模型。
2283
2284### 2. 选择模型服务商
2285
2286选择 WisGate 对应的模型服务商。
2287
2288### 3. 选择使用的模型
2289
2290选择或配置您想使用的模型。
2291
2292### 4. 填写你的 API Key
2293
2294添加 WisGate API Key。
2295
2296如界面需要填写接口地址,请使用 OpenAI 兼容端点:
2297
2298\`\`\`text
2299https://api.wisgate.ai/v1
2300\`\`\`
2301
2302## 三、保存并切换模型
2303
2304配置完成后,保存设置并在自定义模型中选择您刚创建的模型。完毕!现在你就可以在 TRAE 中使用 WisGate 模型进行代码生成、调试、任务分析等工作了。
2305
2306## 四、使用场景
2307
2308- **AI 问答**:围绕当前项目、文件或报错进行技术问答。
2309- **代码生成**:根据需求生成函数、接口、页面和脚本。
2310- **调试修复**:分析异常信息和上下文,给出修复建议。
2311- **任务拆解**:将复杂开发需求拆成可执行步骤。
2312
2313## 常见问题
2314
2315### 无法保存模型配置
2316
2317请检查 API Key 是否有效,并确认接口地址为 \`https://api.wisgate.ai/v1\`。
2318
2319### 生成结果不符合预期
2320
2321优先确认当前选择的是刚配置的 WisGate 模型,并在提示词中补充项目背景、目标文件和期望输出格式。
2322`,mn=`# ZCode
2323
2324ZCode is a full-featured Agentic Development Environment (ADE) designed for Long Horizon Tasks, using its proprietary ZCode Agent, multi-framework compatibility, Bot integration, and mobile remote control.
2325
2326With WisGate, ZCode helps you understand projects, write code, fix issues, and complete complex development tasks more efficiently.
2327
2328## 1. Install ZCode
2329
2330Visit the ZCode website and download the version for your operating system.
2331
2332## 2. Configure WisGate
2333
23341. Click the **Settings** button in the bottom left corner, navigate to **Model Providers** via the left sidebar
23352. Enter your API Key in the provider configuration
23363. For manual configuration, confirm the endpoint URLs:
2337   - OpenAI: \`https://api.wisgate.ai/v1\`
2338   - Anthropic Base URL: \`https://api.wisgate.ai\`
2339   - Anthropic Messages API: \`https://api.wisgate.ai/v1/messages\`
2340
2341If ZCode asks you to choose only one protocol, use the OpenAI-compatible endpoint for general chat and GPT-family models. Use the Anthropic Base URL for Claude-family models or Anthropic Messages-compatible features. Only enter the Messages API URL when the interface asks for the full request URL.
2342
2343## 3. Save and Switch Models
2344
2345After configuration, save your settings and select the configured model in ZCode's model switcher menu.
2346
2347## 4. Use Cases
2348
2349- **Long-horizon development**: Let the agent complete cross-file changes, verification, and summaries step by step.
2350- **Project understanding**: Analyze folder structure, module responsibilities, and business workflows.
2351- **Issue fixing**: Generate fix suggestions from error logs, failed tests, and related code.
2352- **Remote collaboration**: Reuse the same model configuration in Bot or mobile remote-control workflows.
2353
2354## FAQ
2355
2356### Should I choose the OpenAI or Anthropic protocol?
2357
2358It depends on the model and feature. The OpenAI-compatible endpoint is \`https://api.wisgate.ai/v1\`, the Anthropic Base URL is \`https://api.wisgate.ai\`, and the full Messages API URL is \`https://api.wisgate.ai/v1/messages\`. If you are unsure, choose the protocol that matches the model family.
2359
2360### Model calls fail
2361
2362Check that the API key, protocol type, and endpoint URL match, and confirm that the configured model is selected in the model switcher.
2363`,Cn=`# ZCode
2364
2365ZCode 是一款面向 Long Horizon Task 的全功能 Agentic Development Environment(ADE),通过自研 ZCode Agent、多框架兼容、Bot 接入与移动端 Remote 控制,让 AI Agent 端到端稳定完成跨度长、步骤多的开发任务。
2366
2367搭配 WisGate,ZCode 能帮助你更高效地理解项目、编写代码、修复问题和完成复杂开发任务。
2368
2369## 一、安装 ZCode
2370
2371访问 ZCode 官网,下载并安装适合您操作系统的版本。
2372
2373## 二、配置 WisGate
2374
23751. 点击左下角**设置**按钮,通过左侧导航栏进入**模型供应商**页面
23762. 在供应商配置中填写 API Key
23773. 手动配置时,请确认 OpenAI 接口地址和 Anthropic 接口分别为:
2378   - OpenAI: \`https://api.wisgate.ai/v1\`
2379   - Anthropic Base URL: \`https://api.wisgate.ai\`
2380   - Anthropic Messages API: \`https://api.wisgate.ai/v1/messages\`
2381
2382如果 ZCode 只要求选择一种协议,通用聊天和 GPT 系列模型优先使用 OpenAI 兼容地址;Claude 系列或 Anthropic Messages 兼容能力优先使用 Anthropic Base URL。仅当界面要求完整请求地址时,才填写 Messages API 地址。
2383
2384## 三、保存并切换模型
2385
2386配置完成后,保存设置,即可在 ZCode 的模型切换菜单中选择刚配置好的模型。
2387
2388## 四、使用场景
2389
2390- **长任务开发**:让 Agent 按步骤完成跨文件修改、验证和总结。
2391- **项目理解**:分析目录结构、模块职责和业务流程。
2392- **修复问题**:根据错误日志、测试失败信息和相关代码生成修复建议。
2393- **远程协作**:在 Bot 或移动端 Remote 控制场景中复用同一套模型配置。
2394
2395## 常见问题
2396
2397### 应该选择 OpenAI 还是 Anthropic 协议?
2398
2399取决于模型和功能。OpenAI 兼容地址为 \`https://api.wisgate.ai/v1\`,Anthropic Base URL 为 \`https://api.wisgate.ai\`,完整 Messages API 地址为 \`https://api.wisgate.ai/v1/messages\`。如果不确定,先按模型所属协议选择。
2400
2401### 模型调用失败
2402
2403请检查 API Key、协议类型和接口地址是否匹配,并确认已经在模型切换菜单中选择了刚配置的模型。
2404`,fn={key:0,class:"bg-white rounded-2xl border border-slate-200/80 overflow-hidden"},yn={class:"relative px-8 sm:px-10 pt-8 pb-6 border-b border-slate-100 bg-gradient-to-b from-slate-50/50 to-white"},An={class:"flex items-start gap-5"},In={class:"size-12 rounded-2xl flex items-center justify-center shrink-0"},vn={class:"min-w-0 flex-1"},_n={class:"flex items-center gap-3 flex-wrap"},wn={class:"text-xl font-bold text-slate-900 tracking-tight"},bn={class:"text-sm text-slate-500 leading-relaxed"},Pn={class:"px-8 sm:px-10 py-8 sm:py-10"},Gn={class:"integration-md"},kn=y({__name:"[slug]",setup(Wn){const{t,locale:h}=A(),g=k(),m=v(),C=I(),a=r(()=>g.params.slug),n=r(()=>L(a.value));if(!n.value){const i=O[0].slug;m.replace(C(`/developers/integrations/${i}`))}_(r(()=>({title:n.value?`${t(n.value.nameKey)} - ${t("site.integrations.title")}`:t("seo.integrations.title"),meta:[{name:"description",content:n.value?t(n.value.descKey):t("seo.integrations.description")}]})));const s=Object.assign({"/content/integrations/claude-code-ide.en.md":E,"/content/integrations/claude-code-ide.zh.md":T,"/content/integrations/claude-code.en.md":U,"/content/integrations/claude-code.zh.md":x,"/content/integrations/claude-desktop.en.md":M,"/content/integrations/claude-desktop.zh.md":R,"/content/integrations/cline.en.md":j,"/content/integrations/cline.zh.md":K,"/content/integrations/codebuddy.en.md":D,"/content/integrations/codebuddy.zh.md":B,"/content/integrations/codex-cli.en.md":N,"/content/integrations/codex-cli.zh.md":F,"/content/integrations/crush.en.md":H,"/content/integrations/crush.zh.md":z,"/content/integrations/cursor.en.md":J,"/content/integrations/cursor.zh.md":V,"/content/integrations/droid.en.md":Y,"/content/integrations/droid.zh.md":Z,"/content/integrations/goose.en.md":q,"/content/integrations/goose.zh.md":Q,"/content/integrations/hermes.en.md":$,"/content/integrations/hermes.zh.md":X,"/content/integrations/junie.en.md":nn,"/content/integrations/junie.zh.md":en,"/content/integrations/kilo.en.md":tn,"/content/integrations/kilo.zh.md":on,"/content/integrations/lingma.en.md":an,"/content/integrations/lingma.zh.md":sn,"/content/integrations/openclaw.en.md":rn,"/content/integrations/openclaw.zh.md":ln,"/content/integrations/opencode.en.md":dn,"/content/integrations/opencode.zh.md":cn,"/content/integrations/roo.en.md":pn,"/content/integrations/roo.zh.md":un,"/content/integrations/trae.en.md":hn,"/content/integrations/trae.zh.md":gn,"/content/integrations/zco
2404de.en.md":mn,"/content/integrations/zcode.zh.md":Cn}),f=r(()=>{if(!a.value)return"";const i=h.value==="zh"?"zh":"en",l=Object.keys(s).find(o=>o.endsWith(`/${a.value}.${i}.md`));if(!l){const o=Object.keys(s).find(d=>d.endsWith(`/${a.value}.en.md`));return o?s[o]??"":""}return s[l]??""});return(i,l)=>{const o=G,d=S;return n.value?(P(),w("div",fn,[e("div",yn,[e("div",An,[e("div",In,[c(o,{name:n.value.icon,class:"size-12"},null,8,["name"])]),e("div",vn,[e("div",_n,[e("h2",wn,p(u(t)(n.value.nameKey)),1)]),e("p",bn,p(u(t)(n.value.descKey)),1)])])]),e("div",Pn,[e("div",Gn,[c(d,{value:f.value},null,8,["value"])])])])):b("",!0)}}}),Tn=W(kn,[["__scopeId","data-v-b5b51b1b"]]);export{Tn as default};

Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.