# Connect the Context7 MCP server to Unblocked to give AI agents live library documentation URL: https://getunblocked.com/integrations/context7/ Unblocked connects to Context7’s MCP server so AI coding agents can query up-to-date, version-specific library documentation and code examples live — then its Context Engine reconciles what they find with the code, pull requests, and conversations behind your own use of those libraries. --- ## Context7 integration details - **Category:** Documentation - **Live-queried data:** library docs, code examples - **Supported products:** Context7 - **Works with:** Claude Code, Cursor, Devin, GitHub Copilot, Codex, OpenCode, VS Code — via one MCP server connection - **Maintained by:** Context7 · verified by Unblocked - **Content reviewed:** October 20, 2026 - **Website:** ## What the Context7 integration adds Coding agents learn libraries from training data that lags behind real releases, so they suggest APIs that were renamed, deprecated, or never existed in the version you run. Context7’s MCP server gives agents current, version-specific documentation and code examples for the libraries they write against, and Unblocked connects to it once so every agent can look up a library without its own Context7 setup. This is a third-party MCP server: Context7 hosts and maintains it, and Unblocked connects to it as a verified MCP server. Nothing is ingested or indexed — Unblocked calls the read-only tools you select at the moment a question needs them, so documentation reflects what Context7 serves for that library right now. Wiring Context7’s MCP server straight into an agent gives it generic library docs. Unblocked’s Context Engine reasons across those docs and your other sources — following a library’s recommended pattern to the pull request that pinned an older version, the Slack thread about the migration, and the ticket tracking the upgrade — and returns one cited answer that fits your codebase. Once connected, any MCP-compatible coding agent — Claude Code, Cursor, Devin, GitHub Copilot, Codex — reaches Context7 context through a single Unblocked MCP server connection. ### What you get - **Current, version-specific library docs** — Agents pull documentation and code examples for the library they are using, from Context7’s own MCP server, instead of relying on whatever their training data remembers. - **One connection for every agent** — Connect Context7 to Unblocked once. Claude Code, Cursor, Codex, and GitHub Copilot reach it through the Unblocked MCP server, with no per-agent Context7 configuration or API key. - **Read-only lookups you control** — Unblocked only resolves library IDs and queries documentation, using the Context7 tools you select. Change the selection anytime from Settings → Data Sources. - **Library docs reconciled with your code** — The Context Engine weighs Context7 documentation against your code, pull requests, and conversations, so an agent learns both how a library works and how your team decided to use it — with citations. ## How to connect Context7 to Unblocked Prerequisites: - An Unblocked account (free to start). - A Context7 account and API key from the Context7 dashboard, which raises rate limits and covers any private repositories you have added to Context7. - An MCP-compatible coding agent (Claude Code, Cursor, Devin, etc.) to use the context in your editor. ### Step 1: Connect Context7 to Unblocked 1. In Unblocked, go to Settings → Data Sources → Connect another data source, and under Documentation select Context7. 2. The Context7 MCP server URL is filled in for you. Choose HTTP Headers, add an Authorization header with your Context7 API key as Bearer YOUR_API_KEY, then click Connect. 3. Once Unblocked discovers the server’s tools, select the read-only tools it may use and click Save Settings. ### Step 2: Install the Unblocked MCP server On macOS or Linux, one command auto-detects and configures the supported agents on your machine: ```bash curl -fsSL https://getunblocked.com/install-mcp.sh | bash ``` On Windows, or when you prefer remote MCP, add the server manually using the configuration below or follow the manual install guide: Claude Code — add the remote server, then run /mcp in Claude Code, select unblocked, and log in via OAuth: ```bash claude mcp add -s user --transport http unblocked https://getunblocked.com/api/mcpsse ``` Cursor — open View: Open MCP Settings → Add Custom MCP, add the server, then click Connect to authenticate via OAuth: ```json { "mcpServers": { "unblocked": { "url": "https://getunblocked.com/api/mcpsse" } } } ``` Any MCP-compatible client (Devin, Codex, VS Code, Claude Desktop, and others) can connect to the remote server with OAuth: `https://getunblocked.com/api/mcpsse` ## Context7 MCP server tools Context7 hosts and maintains this MCP server. Unblocked calls only the read-only tools you select, at the moment a question needs them: - **`resolve-library-id`** — Converts a library name into a Context7-compatible library ID, ranking matches against the question being asked. - **`query-docs`** — Returns documentation and code examples for a Context7 library ID that answer a specific question, optionally ranking examples in a chosen language first. ## Connect Context7 to Claude Code, Cursor, and other AI agents To give Claude Code, Cursor, GitHub Copilot, Codex, or another MCP-compatible AI coding agent access to Context7 context, connect Context7 to Unblocked and add the single Unblocked MCP server to your agent. Unblocked acts as the MCP context layer for your Context7 data, searching library docs and code examples alongside your other connected sources and returning cited results through these tools: - **`context_research`** — Researches a question across your connected sources and calls the Context7 tools you selected when library documentation is relevant, returning a synthesized answer with citations. - **`context_get_urls`** — Fetches the full content behind URLs your agent encounters across connected sources — Jira, Linear, Notion, Confluence, Slack, GitHub, and more. ## Questions to ask an AI agent about Context7 Questions you (or your agent) can ask Unblocked about your Context7 data: - "What does the current Next.js documentation recommend for caching, and how does our app router code differ from it?" - "We pinned an older version of our ORM — which pull request did that, and what changes in the latest docs would we pick up by upgrading?" - "Is the auth helper our team discussed in Slack still the recommended approach in the library’s docs?" - "Which of our services call the API that the library’s docs now mark as deprecated, and is there a ticket to migrate them?" ## Context7 integration FAQ ### What is the Context7 MCP server? The Context7 MCP server is a remote Model Context Protocol server, hosted and maintained by Context7, that gives AI agents up-to-date, version-specific documentation and code examples for software libraries. Unblocked connects to it as a verified MCP server, so every agent reaches it through one Unblocked connection. ### Which libraries can agents look up through Context7? Any library Context7 has documentation for. Agents first resolve a library name to a Context7 library ID, then query that library’s documentation with a specific question. Libraries from private repositories are available when your Context7 account and API key include them. ### How is Context7 different from the docs Unblocked already connects to? Context7 covers public library and framework documentation maintained outside your company. Unblocked’s own sources cover your internal knowledge — code, pull requests, Confluence, Notion, Slack, and tickets. Together, an agent can see what a library recommends and what your team chose to do with it. ### Can the Context Engine tell when library docs conflict with our code? Yes. When Context7 documentation describes a different API or pattern than the one in your codebase, the Context Engine weighs both and surfaces the pull requests and conversations that explain the gap, rather than letting an agent overwrite a deliberate choice. ### Does Unblocked store the documentation it gets from Context7? No. Nothing from Context7 is ingested or indexed. Unblocked queries Context7 while it answers a question, so agents always see the documentation Context7 serves at that moment. ## Related integrations - [GitHub](https://getunblocked.com/integrations/github/) — Repositories, pull requests, issues, discussions, and full code history — indexed, kept in sync on every push, and cited in answers. - [Mintlify](https://getunblocked.com/integrations/mintlify/) — Docs project pages, comments, and analytics from Mintlify’s verified MCP server, queried live while Unblocked answers — nothing is ingested or indexed. - [External websites](https://getunblocked.com/integrations/public-websites/) — Any public web page you specify — vendor docs, API references, standards — crawled and cited alongside your internal sources. - [Stack Overflow for Teams](https://getunblocked.com/integrations/stackoverflow-teams/) — Internal questions, answers, and articles from Stack Overflow for Teams, so hard-won troubleshooting knowledge shows up in answers. ## Get started with Unblocked - Start a free trial: - Pricing: - All integrations: