# Context Repo

> Context Repo is the AI Context Repo for agents. Save prompts and
> documents with version history, organize them in collections, and
> access them from ChatGPT, Claude, Cursor, and 90+ MCP-compatible
> clients via a 29-tool MCP server (27 power tools + OpenAI Apps SDK
> search/fetch).

- Prompt templates with `${variable}` placeholder syntax (clients
  interpolate at call time; templates are stored verbatim)
- Semantic search across all content (1536-dimension vectors)
- Hierarchical Deep Search inside document content (chunk-level)
- Version history with full diff tracking on prompts and documents
- 75+ file format processing via LlamaIndex Cloud
- Web scraping via Firecrawl integration
- 30 REST API operations under /v1 with Bearer JWT or API-key auth

## When to use Context Repo

Use Context Repo when an agent needs to (a) retrieve a versioned prompt
template, (b) search across a user's stored prompts/documents/collections,
(c) persist a new prompt or document for later reuse, or (d) sync
context across Claude, Cursor, ChatGPT, VS Code, Windsurf, Factory, or
Amp. Do NOT use Context Repo as a general web search engine, a
high-volume vector-DB substitute, a multi-tenant team workspace, or an
LLM conversation history store.

## Getting Started

- [MCP Server Setup](https://contextrepo.com/mcp-server): Connect Context Repo to Claude, Cursor, or any MCP client in under 60 seconds
- [Quickstart](https://contextrepo.com/docs/getting-started/quickstart): 60-second integration walkthrough — sign up, create your first prompt, connect an AI assistant
- [ChatGPT App](https://contextrepo.com/chatgpt-app): Native ChatGPT integration via OAuth
- [Chrome Extension](https://contextrepo.com/chrome-extension): Save content from any webpage directly to Context Repo
- [User Documentation](https://contextrepo.com/docs): Setup guides, dashboard reference, MCP and REST integration docs, and FAQ for Context Repo across 90+ AI clients
- [Ask the Docs](https://ask.contextrepo.com): Conversational AI that answers questions from the Context Repo documentation
- [Full Agent Context (`llms-full.txt`)](https://contextrepo.com/llms-full.txt): Single-fetch knowledge pack for agents

## Features

- [Document Management](https://contextrepo.com/agent/features/documents): Store knowledge artifacts with automatic embedding generation for 75+ file formats
- [Semantic Search](https://contextrepo.com/agent/features/search): Vector and keyword search across all prompts, documents, and collections
- [Deep Search](https://contextrepo.com/agent/features/deep-search): Hierarchical navigation inside document content with chunk-level precision
- [Collections](https://contextrepo.com/agent/features/collections): Organize content into named folders with scoped semantic search and per-collection MCP access from any AI client
- [Prompt Management](https://contextrepo.com/agent/features/prompts): Create, version, and share prompt templates with placeholder syntax
- [MCP Server](https://contextrepo.com/mcp-server): 29 tools (27 power tools + Apps SDK search/fetch) with OAuth and API key authentication, including `reason` for synthesized, cited answers across your documents
- [REST API](https://contextrepo.com/agent/features/api): 30 HTTP operations under /v1 with Bearer JWT and API key auth

## Developer Resources (HTML reference)

- [REST API Documentation](https://contextrepo.com/docs/api): Human-readable REST API reference — base URL derivation, response envelope, error format, CORS policy, rate limits
- [REST API Authentication](https://contextrepo.com/docs/api/authentication): Bearer JWT (Clerk) and API-Key setup with example requests
- [REST API Endpoints](https://contextrepo.com/docs/api/endpoints): Full reference for all 30 REST endpoints with request/response examples
- [Webhooks](https://contextrepo.com/docs/webhooks): Real-time event delivery with Svix signature verification, payload schema, retry behavior, and idempotency guidance

## API Specification

- [OpenAPI 3.1 (JSON)](https://contextrepo.com/openapi.json): Machine-readable spec for the 30 REST operations
- [OpenAPI 3.1 (YAML)](https://contextrepo.com/api/openapi.yaml): YAML mirror of the OpenAPI spec
- [API Catalog (RFC 9727)](https://contextrepo.com/.well-known/api-catalog): Linkset of REST artifacts
- [Markdown homepage (`index.md`)](https://contextrepo.com/index.md): Plain-markdown homepage rendering
- [Markdown pricing (`pricing.md`)](https://contextrepo.com/pricing.md): Plain-markdown pricing reference

## MCP Server

- [MCP Endpoint](https://contextrepo.com/mcp): Streamable HTTP transport. Clients that send `Accept: text/event-stream` receive SSE-framed JSON-RPC; JSON-only clients receive a single chunked response (an Accept-shim unframes the SSE on the wire automatically).
- [MCP Discovery](https://contextrepo.com/.well-known/mcp.json): MCP server capabilities and connection info (SEP-1649)
- [MCP Server Card](https://contextrepo.com/.well-known/mcp/server-card.json): All 29 tool registrations with descriptions
- [Agent Catalog](https://contextrepo.com/agent): Machine-readable capability catalog (JSON or markdown via Accept)

## Agent Discovery

- [OpenAI ai-plugin manifest](https://contextrepo.com/.well-known/agent.json): Agent discovery file
- [A2A Agent Card](https://contextrepo.com/.well-known/agent-card.json): Agent-to-Agent skills catalog
- [OAuth Authorization Server (RFC 8414)](https://contextrepo.com/.well-known/oauth-authorization-server): Clerk-backed authorization-server metadata
- [OAuth Protected Resource — REST (RFC 9728)](https://contextrepo.com/.well-known/oauth-protected-resource/v1): Scopes for the /v1 REST API
- [OAuth Protected Resource — MCP (RFC 9728)](https://contextrepo.com/.well-known/oauth-protected-resource/mcp): OAuth scopes for the MCP transport

## Resources

In-depth articles on AI context management, MCP servers, prompt
versioning, semantic search, and integrations across Claude, Cursor,
and ChatGPT.

- [What is an AI Context Repo for Agents?](https://contextrepo.com/resources/what-is-an-ai-context-repo-for-agents): What an AI context repository is, what belongs inside it, and how Context Repo gives humans and agents one home for prompts, documents, and collections across every AI tool
- [Prompt and Document Management for AI Agents](https://contextrepo.com/resources/prompt-and-document-management-for-ai-agents): How a context repository handles version history, variable interpolation, semantic search, and 75+ file formats over MCP and REST
- [How MCP Servers Connect AI Agents to Knowledge Bases](https://contextrepo.com/resources/how-mcp-servers-connect-ai-agents-to-knowledge-bases): What an MCP server is, what Context Repo's 29-tool server exposes, and how it connects Claude, Cursor, ChatGPT, and 90+ AI clients to your context repository
- [Semantic Search and Deep Search: Two Retrieval Layers](https://contextrepo.com/resources/semantic-search-and-deep-search-two-retrieval-layers): Vector similarity at the catalog level, hierarchical chunk navigation inside documents, and how to pick the right retrieval surface for an agent's question
- [Using Context Repo with Claude, Cursor, and ChatGPT](https://contextrepo.com/resources/using-context-repo-with-claude-cursor-and-chatgpt): How a single context repository travels across Cursor, Claude Desktop, Claude.ai, and ChatGPT, with workflows for the engineer, the researcher, and the operator

Resources index: https://contextrepo.com/resources/llms.txt
Resources RSS: https://contextrepo.com/resources/feed.xml

## Modular Sections

- [Documentation index](https://contextrepo.com/docs/llms.txt): Section-scoped llms.txt for the user-documentation surface
- [REST API index](https://contextrepo.com/api/llms.txt): Section-scoped llms.txt for the /v1 REST API surface
- [Developers index](https://contextrepo.com/developers/llms.txt): Section-scoped llms.txt for agent / MCP / SDK integrators
- [Resources index](https://contextrepo.com/resources/llms.txt): Section-scoped llms.txt for long-form articles

## Constraints & Rate Limits

- Rate limits (per authenticated user, sliding window): scrape 10/60s, api 100/60s, readonly 120/60s. Headers `X-RateLimit-*` and `Retry-After` are emitted on 429 only. Fail-open if Upstash Redis is unavailable.
- Vector dimension: 1536 (text-embedding-3-small). Mismatched embeddings fail silently.
- Document body cap: 2 MB per content field on the wire.
- Pagination: opaque base64 cursors; default page size 20, max 100.
- DELETE responses are 200 + `{success, id}` (not 204) for MCP-client compatibility.

## Pricing

- [Plans & Pricing](https://contextrepo.com/pricing): 7-day free trial available on the Pro and Pro Max plans. Pro ($9.99/mo) includes 29 MCP tools (27 power tools + Apps SDK search/fetch), ChatGPT App, Chrome Extension, version history. Pro Max ($19.99/mo) at 2× storage, scrapes, and uploads. Ultimate ($99.99/mo) at 10×.

## SDK & packages

- [npm: context-repo-mcp](https://www.npmjs.com/package/context-repo-mcp): Official MCP server package (`npx context-repo-mcp`)
- [GitHub: Gitmaxd/context-repo-mcp](https://github.com/Gitmaxd/context-repo-mcp): MCP server source code
- [Chrome Web Store](https://chromewebstore.google.com/detail/context-repo/hlmpbcbphnhecbcakbnlobgighfnlafi): Context Repo Chrome extension

## Contact

- Email: support@contextrepo.com
- X / Twitter: https://x.com/contextrepo
- GitHub: https://github.com/Gitmaxd/context-repo-mcp
- Status: https://contextrepo.com/status
- About: https://contextrepo.com/about
- Public agent rules: https://contextrepo.com/AGENTS.md

## Optional

- [Privacy Policy](https://contextrepo.com/policies/privacy)
- [Terms of Service](https://contextrepo.com/policies)