A CLI code search tool that parses code structure and ranks results by relevance, with no index required.
sourcebot-dev/sourcebot
Self-hosted code context layer that indexes repos across hosts so developers and AI agents can search and navigate them.
Open-source alternative to:Sourcegraph
- Stars
- 4,003
- Forks
- 372
- Open issues
- 96
- Last commit
- 2 days ago
What is it?
Sourcebot is a self-hosted tool for indexing code from many repositories in one place. Developers can search across all indexed repos and branches without cloning each one, using regular expressions and boolean queries filtered by repo, file, language, or symbol definition.
It also exposes an MCP server, so AI agents such as Claude Code, Codex, or Cursor can search code, read files, and look up symbol definitions and references in repos that are not checked out locally. Repositories can be connected from GitHub, GitLab, Bitbucket, Azure DevOps, or local Git directories, with optional permission syncing based on code host access rules.
An Ask feature answers natural-language questions about the indexed code, citing source files, and runs on a language model provider the user configures. Setup is handled by an npx wizard, Docker Compose, or a Helm chart.
Key features
- 01Searches across all indexed repos and branches with regex and boolean queries.
- 02Provides an MCP server for agents to search code, read files, and find symbol references.
- 03Connects to GitHub, GitLab, Bitbucket, Azure DevOps, and local Git directories.
- 04Answers codebase questions with cited source files using a user-configured language model.
- 05Supports permission syncing based on code host repository permissions.
Use cases
- Finding callers of an API across many microservices without cloning each repo.
- Giving an AI coding agent context about dependencies that are not checked out locally.
- Searching a company's codebase in one web interface with consistent access rules.
- Asking natural-language questions about how a feature works in an unfamiliar codebase.
Quick start
- 1
Run the setup wizard (needs Node.js, Docker, and Docker Compose)
npx setup-sourcebot
Things to know
- Requires Node.js, Docker, and Docker Compose to be installed.
- Ask Sourcebot requires your own language model provider API key (Bring Your Own Key).
Topics
FAQ
- What is sourcebot?
- Self-hosted code context layer that indexes repos across hosts so developers and AI agents can search and navigate them. Sourcebot is a self-hosted tool for indexing code from many repositories in one place. Developers can search across all indexed repos and branches without cloning each one, using regular expressions and boolean queries filtered by repo, file, language, or symbol definition.
- Is sourcebot free to use? What license does it have?
- sourcebot is open source under the MIT license. Check the license terms before using it in commercial products.
- Is sourcebot still maintained?
- Actively maintained. The last commit to sourcebot was on Oct 20, 2026.
- What can I use sourcebot for?
- Finding callers of an API across many microservices without cloning each repo. Giving an AI coding agent context about dependencies that are not checked out locally. Searching a company's codebase in one web interface with consistent access rules. Asking natural-language questions about how a feature works in an unfamiliar codebase.
- How do I install sourcebot?
- Run the setup wizard (needs Node.js, Docker, and Docker Compose): npx setup-sourcebot
- Where can I get sourcebot?
- The source code and installation instructions are on GitHub: https://github.com/sourcebot-dev/sourcebot
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Indexes a codebase into a knowledge graph of dependencies, call chains and execution flows, and serves it to AI coding agents via MCP tools.
Codebase intelligence for developers and coding agents: call graph, git history, health scores, dead code and docs, served over MCP and a local dashboard.
A zero-dependency C++23 CLI and MCP server that gives coding agents a ranked map of a codebase and checks their edits for blast radius, tests to run and quality regressions.
A CLI-driven context graph that indexes your codebase, issues, and team knowledge so AI coding agents work with project-specific context.
An MCP server that indexes code, docs and infrastructure into a local knowledge graph so AI coding agents can query structure instead of re-reading files.