Context7
Up-to-date docs for AI agents
作者提供的说明(英文原文)
Search Christian Perez's blog, read posts, and ask about his work, podcast, and writing.
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
适用于网页版、桌面版和手机版 Claude:
https://mcp.thechrisgrey.com在终端中运行这条命令
claude mcp add --transport http thechrisgrey https://mcp.thechrisgrey.com粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"thechrisgrey": {
"url": "https://mcp.thechrisgrey.com"
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"thechrisgrey": {
"type": "http",
"url": "https://mcp.thechrisgrey.com"
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
ChatGPT 只能连接互联网上的服务器:
https://mcp.thechrisgrey.com菜单名称可能因套餐和版本不同而略有差异。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
我们还没有读取这个服务器的工具列表;我们会轮流检查各个端点。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月7日
注册名称已通过域名 thechrisgrey.com 验证(注册中心会核实域名所有权)
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
尚未检查;我们会轮流检查各个端点
尚未检查;我们会轮流检查各个端点
尚未检查;我们会在检查端点时读取工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
它在作者的云端运行。通过互联网连接,无需安装,但你发送的内容会经过作者的服务器。
GitHub 仓库最后更新于–,最新版本(1.0.0)于7个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,所以接受远程 MCP 服务器的应用也可以使用,例如开启开发者模式的 ChatGPT。
同一分类中有:Context7, Codebase Memory, Exa
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Up-to-date docs for AI agents
Codebase knowledge graph for AI agents — 162 languages, sub-ms queries, 99% fewer tokens.
Connect AI agents to Exa for web search, content fetching, and multi-step research.
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
Personal website for Christian Perez (@thechrisgrey) | Founder & CEO of Altivum Inc. | Former Green Beret (18D Special Forces Medical Sergeant) | Bronze Star Recipient | Host of The Vector Podcast | Author of "Beyond the Assessment"
[https://thechrisgrey.com](https://thechrisgrey.com)
| Layer | Technology |
|---|---|
| Frontend | React 19, TypeScript 5, Vite 5 (Vitest 4, ESLint 10 flat config) |
| Styling | Tailwind CSS 4 (CSS-first @theme config, Oxide engine) |
| Routing | React Router 7 |
| CMS | Sanity.io (blog content) |
| AI Chat | Amazon Bedrock (Claude Haiku 4.5) + RAG |
| Auth | Amazon Cognito (admin panel) |
| Hosting | AWS Amplify |
| CDN/Analytics | Cloudflare Web Analytics |
| Testing | Vitest + React Testing Library + Cypress |
| Code Quality | ESLint, strict TypeScript |
┌─────────────────────────────────────────────────────────────────┐
│ FRONTEND (React) │ │ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │ │ │ Home │ │ Podcast │ │ Blog │ │ Chat │ ... │ │ └──────────┘ └──────────┘ └──────────┘ └──────────┘ │ └─────────────────────────────────────────────────────────────────┘ │ │ │ │ ▼ ▼ ▼ ▼ ┌───────────┐ ┌───────────┐ ┌───────────┐ ┌───────────────┐ │ Amplify │ │ YouTube │ │ Sanity │ │ Bedrock │ │ Hosting │ │ Data API │ │ CMS │ │ + RAG (KB) │ └───────────┘ └───────────┘ └───────────┘ └───────────────┘ (build-time) (runtime) (Lambda stream)
User Message
│ ▼ Lambda Function URL (streaming) │ ├──> Bedrock Knowledge Base (retrieve context) │ │ │ ▼ │ S3 Vectors (autobiography chunks) │ ├──> Bedrock Guardrails (content filtering) │ └──> Bedrock ConverseStream API (Claude Haiku 4.5) │ ▼ Streaming Response -> Client
| Service | Resource | Region | Purpose |
|---|---|---|---|
| Amplify | d3du8eg39a9peo | us-east-2 | Hosting, CI/CD |
| Lambda | thechrisgrey-chat-stream | us-east-1 | AI chat streaming |
| Lambda | thechrisgrey-contact-email | us-east-2 | Contact form |
| Lambda | thechrisgrey-newsletter-subscription | us-east-2 | Newsletter |
| Lambda | thechrisgrey-kb-sync | us-east-1 | KB auto-sync on S3 changes |
| Lambda | thechrisgrey-kb-builder | us-east-1 | KB admin CRUD operations |
| Lambda | thechrisgrey-metrics | us-east-1 | Metrics collection and reporting |
| Bedrock KB | ARFYABW8HP | us-east-1 | RAG knowledge base |
| Cognito | User Pool | us-east-1 | Admin authentication |
| S3 | thechrisgrey-kb-source | us-east-1 | KB source documents |
| S3 | thechrisgrey-vectors | us-east-1 | Vector embeddings |
| DynamoDB | thechrisgrey-chat-ratelimit | us-east-1 | Rate limiting (20 req/hr/IP) |
| Secrets Manager | thechrisgrey/youtube-api-key | us-east-1 | YouTube API key |
| CloudWatch | tcg-AI-chat | us-east-1 | Bedrock invocation logs |
.nvmrc)git clone https://github.com/thechrisgrey/thechrisgrey.git
cd thechrisgrey npm install
npm run dev # Start dev server at http://localhost:5173
npm run build # Full production build npm run preview # Preview production build npm run lint # Run ESLint
The production build runs these steps in sequence:
validate-env -> podcast-episodes -> lint -> tsc -> vite build -> sitemap -> rss
| Framework | Scope | Command |
|---|---|---|
| Vitest | Unit + Integration | npm test |
| Vitest | Watch mode | npm run test:watch |
| Vitest | Coverage report | npm run test:coverage |
| Cypress | E2E (headless) | npm run cy:run |
| Cypress | E2E (interactive) | npm run cy:open |
| node:test | Lambda unit + contract | npm run test:lambda |
The test suite includes unit tests for components, hooks, and utilities; integration tests for page-level behavior; and end-to-end tests covering navigation, chat, blog, and contact flows.
Critical external boundaries — Cognito admin auth, the live Sanity client, Bedrock KB retrieval, and the Bedrock guardrail — are also covered by opt-in contract tests that exercise the real services (green mocks alone can drift from reality). They skip cleanly during npm run test:lambda and CI, and only run when their flag + AWS/Sanity credentials are present:
COGNITO_CONTRACT_TESTS=1 COGNITO_CLIENT_ID=… COGNITO_TEST_EMAIL=… COGNITO_TEST_PASSWORD=… \
node --test lambda/shared/tests/cognito-contract.test.mjs SANITY_CONTRACT_TESTS=1 node --test lambda/chat-stream/tests/sanity-contract.test.mjs KB_RETRIEVE_CONTRACT_TESTS=1 node --test lambda/chat-stream/tests/kb-retrieve-contract.test.mjs BEDROCK_CONTRACT_TESTS=1 node --test lambda/blueprint/tests/guardrail-contract.test.mjs
See .env.example for the full list of flags and required permissions. Run these before a release (or on a scheduled job) to catch response-shape/schema drift that unit mocks can't.
Set in AWS Amplify console or in a local .env file:
| Variable | Description | Required |
|---|---|---|
VITE_CHAT_ENDPOINT | Lambda Function URL for AI chat | Yes |
VITE_CONTACT_ENDPOINT | Lambda Function URL for contact form | Yes |
VITE_NEWSLETTER_ENDPOINT | Lambda Function URL for newsletter | Yes |
VITE_COGNITO_USER_POOL_ID | Cognito User Pool ID (admin auth) | Yes |
VITE_COGNITO_CLIENT_ID | Cognito App Client ID | Yes |
VITE_KB_BUILDER_ENDPOINT | Lambda Function URL for KB admin | Yes |
VITE_CHAT_SIGNING_KEY | HMAC signing key for chat requests | Yes |
VITE_METRICS_ENDPOINT | Lambda Function URL for metrics | Yes |
YOUTUBE_API_KEY | YouTube Data API v3 key (build-time only) | Yes |
VITE_ prefixed variables are bundled into the frontend. YOUTUBE_API_KEY is only used during the build step.
Push to main triggers automatic Amplify deployment:
git push origin main
aws amplify start-job \
--app-id d3btl94rm91y13 \ --branch-name main \ --job-type RELEASE \ --region us-east-2 \ --profile personal-dev # Delta Centric 512880383078, the site's home
Each Lambda function is deployed independently via a single verified script. The script installs from the lockfile, dereferences the lambda-shared symlink into the bundle, and verifies the entire module graph resolves before upload — so a deploy can no longer ship a Lambda that crashes on cold start.
# Verify the bundle builds and resolves, without uploading:
npm run deploy:lambda -- chat-stream --dry-run
只显示了开头部分;其余内容请在仓库中阅读