Sourcegraph MCP Server

Connect AI agents and applications to your Sourcegraph instance's code search and analysis capabilities.

Supported on Enterprise plans.

This feature is experimental and might change or be removed in the future.

The Sourcegraph Model Context Protocol (MCP) Server provides AI agents and applications with programmatic access to your Sourcegraph instance's code search, navigation, and analysis capabilities through a standardized interface.

Server Endpoints

The MCP server provides different tool suites for various use cases:

EndpointDescription
/.api/mcpFull suite of Sourcegraph tools
/.api/mcp/deepsearchDeep Search subagent

Example URL:

SHELL
https://your-sourcegraph-instance.com/.api/mcp

Authentication

The Sourcegraph MCP server supports two authentication methods:

OAuth 2.0 with Dynamic Client Registration

Sourcegraph supports MCP spec-compliant OAuth with Dynamic Client Registration (RFC 7591), allowing AI agents to authenticate without pre-configured client IDs.

Enabling Dynamic Client Registration

Enable this feature by setting the following site configuration:

JSON
{ "auth.idpDynamicClientRegistrationEnabled": true }

Once enabled, MCP clients that support OAuth can authenticate automatically. For example, for Amp, you can run:

BASH
amp mcp add sg https://sourcegraph.example.com/.api/mcp

For Claude Code, you can run:

BASH
claude mcp add --transport http sg https://sourcegraph.example.com/.api/mcp

This works similarly for other MCP-compatible agents.

Manual OAuth Setup

If your agent doesn't support Dynamic Client Registration, you can manually create an OAuth application:

  1. Create an OAuth application in your Sourcegraph instance following the instructions here. (Note: you will need the user:all scope)
  2. Use mcp-remote as a fallback with the following configuration:
JSON
{ "sourcegraph": { "type": "stdio", "command": "npx", "args": [ "mcp-remote", "https://your-sourcegraph-instance.com/.api/mcp", "3334", "--static-oauth-client-info", "{\"client_id\":\"<your-client-id>\"}", "--static-oauth-client-metadata", "{\"scope\":\"user:all\"}" ] } }
  • Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and <your-client-id> with the ID of the client you registered.
  • When using the mcp-remote fallback, ensure the OAuth client has one of its redirect URIs set to http://localhost:3334.

Authorization Header

Include your token in the Authorization header:

SHELL
Authorization: token YOUR_ACCESS_TOKEN

Client Integration

The Sourcegraph MCP server can be integrated with various AI tools and IDEs that support the Model Context Protocol.

Amp

To create an access token, visit Creating an access token.

You can add the Sourcegraph MCP server to Amp in two ways:

Option 1: VSCode settings.json

  1. Open VSCode's settings.json file.

  2. Add the following configuration:

    JSON
    { "amp.mcpServers": { "sourcegraph": { "url": "https://your-sourcegraph-instance.com/.api/mcp", "headers": { "Authorization": "token YOUR_ACCESS_TOKEN" } } } }

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

  3. Save the configuration file.

  4. Restart VS Code to apply the new configuration.

Option 2: Amp CLI

Run the following command in your terminal:

BASH
amp mcp add sourcegraph --header "Authorization=token YOUR_ACCESS_TOKEN" https://sourcegraph.sourcegraph.com/.api/mcp

Replace sourcegraph.sourcegraph.com with your Sourcegraph instance URL and set YOUR_ACCESS_TOKEN environment variable to your access token.

Claude Code

You can add the Sourcegraph MCP server to Claude Code in two ways:

Option 1: Project-scoped server (via .mcp.json file)

  1. Create a .mcp.json file in your project root if it doesn't exist.

  2. Add the following configuration:

    JSON
    { "mcpServers": { "sourcegraph": { "type": "http", "url": "https://your-sourcegraph-instance.com/.api/mcp", "headers": { "Authorization": "token YOUR_ACCESS_TOKEN" } } } }

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

  3. Save the configuration file.

  4. Restart Claude Code to apply the new configuration.

Option 2: Locally-scoped server (via CLI command)

You can also add the Sourcegraph MCP server as a locally-scoped server, which is only available to you in the current project:

  1. Run the following command in your terminal:

    BASH
    claude mcp add --transport http sourcegraph https://your-sourcegraph-instance.com/.api/mcp \ --header "Authorization: token YOUR_ACCESS_TOKEN"

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

Locally-scoped servers take precedence over project-scoped servers with the same name and are stored in your project-specific user settings.

Google Gemini Code Assist

You can add the Sourcegraph MCP server to Google Gemini Code Assist by configuring the .gemini/settings.json file:

  1. Open or create the configuration file at ~/.gemini/settings.json (or the equivalent path on your system).

  2. Add the following configuration:

    JSON
    { "mcpServers": { "sourcegraph": { "httpUrl": "https://your-sourcegraph-instance.com/.api/mcp", "headers": { "Authorization": "token YOUR_ACCESS_TOKEN" } } } }

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

  3. Save the configuration file.

  4. Restart Gemini Code Assist to apply the new configuration.

Cursor

You can add the Sourcegraph MCP server to Cursor by configuring it in your MCP settings file:

  1. Open or create the MCP configuration file at ~/.cursor/mcp.json (or the equivalent path on your system).

  2. Add the following configuration:

    JSON
    { "mcpServers": { "sourcegraph": { "type": "http", "url": "https://your-sourcegraph-instance.com/.api/mcp", "headers": { "Authorization": "token YOUR_ACCESS_TOKEN" } } } }

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

  3. Save the configuration file.

  4. Restart Cursor to apply the new configuration.

VS Code

  1. Create ⁠.vscode/mcp.json in your project.

  2. Add the following:

    JSON
    { "servers": { "sourcegraph": { "url": "https://your-sourcegraph-instance.com/.api/mcp", "type": "http", "headers": { "Authorization": "token YOUR_ACCESS_TOKEN" } } }, "inputs": [] }

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

Antigravity

  1. Create ⁠.vscode/mcp.json (Antigravity uses .vscode for configs) in your project.

  2. Add the following:

    JSON
    { "servers": { "sourcegraph": { "url": "https://your-sourcegraph-instance.com/.api/mcp", "type": "http", "headers": { "Authorization": "token YOUR_ACCESS_TOKEN" } } }, "inputs": [] }

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

OpenCode

You can add the Sourcegraph MCP server to OpenCode by configuring it in your MCP settings file:

  1. Open or create the MCP configuration file at ~/.config/opencode/opencode.jsonc (or the equivalent path on your system).

  2. Add the following configuration:

    JSON
    { "mcp": { "sourcegraph": { "type": "remote", "url": "https://your-sourcegraph-instance.com/.api/mcp", "oauth": false, "headers": { "Authorization": "token {env:YOUR_ACCESS_TOKEN}" } } }, "$schema": "https://opencode.ai/config.json" }

    Replace your-sourcegraph-instance.com with your Sourcegraph instance URL and YOUR_ACCESS_TOKEN with your access token.

  3. Save the configuration file.

  4. Restart OpenCode to apply the changes.

Available Tools

All MCP tools implement result limits to ensure efficient operation and prevent context window overflow. These limits are designed to return the most relevant results while maintaining optimal performance. For large result sets, use pagination parameters (after/before cursors) where available, or refine your search with more specific filters.

The MCP server provides these tools for code exploration and analysis:

File & Repository Operations

read_file

Read file contents with line numbers and support for specific ranges and revisions.

Parameters:

  • repo - Repository name (required)
  • path - File path within repository (required)
  • startLine - Starting line number (optional)
  • endLine - Ending line number (optional)
  • revision - Branch, tag, or commit hash (optional)

Use cases: Reading specific files, examining code sections, reviewing different versions

File size limit is 128KB. Use line ranges for larger files.

list_files

List files and directories in a repository path.

Parameters:

  • repo - Repository name (required)
  • path - Directory path (optional, defaults to root)
  • revision - Branch, tag, or commit hash (optional)

list_repos

Search and list repositories by name patterns with pagination support.

Parameters:

  • query - Search pattern for repository names (required)
  • limit - Maximum results per page (optional, default 50)
  • after/before - Pagination cursors (optional)

Perform exact keyword searches with boolean operators and filters.

Parameters:

  • query - Search query with optional filters (required)

Supported filters:

  • repo: - limit to specific repositories
  • file: - search specific file patterns
  • rev: - search specific revisions

Features: Boolean AND/OR operators, regex patterns

Semantic search with flexible linguistic matching.

Parameters:

  • query - Natural language search query (required)

Supported filters:

  • repo: - limit to specific repositories
  • file: - search specific file patterns
  • rev: - search specific revisions

Features: Flexible linguistic matching, stemming, broader results than keyword search

Code Navigation

go_to_definition

Find the definition of a symbol from a usage location.

Parameters:

  • repo - Repository name (required)
  • path - File path containing symbol usage (required)
  • symbol - Symbol name to find definition for (required)
  • revision - Branch, tag, or commit hash (optional)

Features: Cross-repository support, compiler-level accuracy

find_references

Find all references to a symbol from its definition location.

Parameters:

  • repo - Repository name (required)
  • path - File path containing symbol definition (required)
  • symbol - Symbol name to find references for (required)
  • revision - Branch, tag, or commit hash (optional)

Version Control & History

Search commits by message, author, content, files, and date ranges.

Parameters:

  • repos - Array of repository names (required)
  • messageTerms - Terms to search in commit messages (optional)
  • authors - Filter by commit authors (optional)
  • contentTerms - Search in actual code changes (optional)
  • files - Filter by file paths (optional)
  • after/before - Date range filters (optional)

Search actual code changes for specific patterns across repositories.

Parameters:

  • pattern - Search pattern for code changes (required)
  • repos - Array of repository names (required)
  • added - Search only added code (optional)
  • removed - Search only removed code (optional)
  • author - Filter by author (optional)
  • after/before - Date range filters (optional)

compare_revisions

Compare changes between two specific revisions.

Parameters:

  • repo - Repository name (required)
  • base - Base revision (older version) (required)
  • head - Head revision (newer version) (required)
  • first - Maximum file diffs to return (optional, default 50)
  • after - Pagination cursor (optional)

get_contributor_repos

Find repositories where a contributor has made commits.

Parameters:

  • author - Author name or email (required)
  • limit - Maximum repositories to return (optional, default 20)
  • minCommits - Minimum commits required (optional, default 1)

deepsearch_read

Read a Deep Search conversation and return the markdown content of the questions and answers.

Parameters:

  • identifier - The Deep Search conversation identifier (required). Can be:
    • A full Sourcegraph Deep Search URL (e.g., https://sourcegraph.com/deepsearch/abc123 or https://sourcegraph.com/deepsearch/shared/abc123)
    • A read token (e.g., abc123-def456-...)

Use cases: Reading or re-opening Deep Search results, summarizing existing answers, using past Deep Search as context for new questions

Usage Examples

Finding Authentication Code

JSON
{ "method": "tools/call", "params": { "name": "nls_search", "arguments": { "query": "authentication login user" } } }

Reading a Specific File

JSON
{ "method": "tools/call", "params": { "name": "read_file", "arguments": { "repo": "github.com/myorg/myrepo", "path": "src/auth/login.go", "startLine": 1, "endLine": 50 } } }

Finding Recent Changes

JSON
{ "method": "tools/call", "params": { "name": "commit_search", "arguments": { "repos": ["github.com/myorg/myrepo"], "messageTerms": ["bug fix"], "after": "1 week ago" } } }

Best Practices

  1. Repository Scoping: Use list_repos first to find relevant repositories for better performance
  2. Progressive Search: Start with broad searches (nls_search) then narrow with specific tools
  3. File Verification: Use list_files before read_file to verify file existence
  4. Pagination: Use after/before cursors for large result sets
  5. Tool Combinations: Chain tools together (e.g., list_reposcommit_search)
Previous
Deep Search API