# GitHits

> GitHits gives AI coding agents version-aware access to real open-source examples, dependency source code, docs, and package metadata, so they navigate the actual code running in your stack instead of guessing.

AI coding agents can read your local repository, but the open-source stack underneath it — frameworks, SDKs, APIs, infrastructure tooling, dependency internals, and version-specific behavior spread across thousands of repositories — stays invisible. When context is missing, agents loop instead of converge: they retry, rewrite, and improvise, burning tokens and producing fragile integrations. GitHits closes that gap by grounding agents in proven implementations from open source, so they stop guessing and keep moving forward.

GitHits indexes open-source code — from package dependencies and public repositories — into a version-pinned code graph, surfacing source-grounded, distilled results rather than raw repository search.

GitHits ships as a CLI and MCP server that give agents tools for agentic code search, dependency/package inspection, documentation lookup, and grounded open-source examples. Connect an agent with a single command — `npx githits@latest init` — which detects your installed agents, configures them to use GitHits, and signs you in.

## Get started

- [Web app](https://app.githits.com): Sign in and use GitHits in the browser.
- [Install the CLI](https://githits.com/blog/connect-your-agents-to-githits-with-a-single-command/): Run `npx githits@latest init` to detect your installed agents, configure them to use GitHits, and sign in (at the user or project level).

## Documentation

- [Documentation](https://docs.githits.com/): Product documentation, CLI reference, MCP setup guides, and tool reference.
- [Changelog](https://docs.githits.com/changelog/overview): Release notes and product updates.

## Core pages

- [Homepage](https://githits.com/): Open-source code as context for AI coding agents.
- [The Problem](https://githits.com/the-problem/): Why agents get stuck — missing visibility into the open-source stack leads to retry loops, fragile systems, and token burn.
- [The Solution](https://githits.com/the-solution/): Version-aware access to real open-source code, docs, and dependency context so agents inspect the rest of the stack, not just your repo.
- [Why GitHits](https://githits.com/why-githits/): The context layer for the modern tech stack — a discovery engine, not a search engine.
- [The Builders](https://githits.com/the-builders/): The team building GitHits — engineers who rely on AI coding agents daily.
- [FAQ](https://githits.com/faq/): Who GitHits is for, how it works alongside your agent, how the index works, private-repo and data privacy guarantees, and open-source license handling.

## Blog

- [Blog index](https://githits.com/blog/): Build notes and product thinking on agents, context, and how we build GitHits.
- [GitHits CLI Launch](https://githits.com/blog/githits-cli/): 14 tools for agentic code search, package inspection, documentation, and grounded open-source examples.
- [Connect Your Agents with a Single Command](https://githits.com/blog/connect-your-agents-to-githits-with-a-single-command/): One init command detects installed agents, configures them to use GitHits, and signs in at the user or project level.
- [Code Navigation](https://githits.com/blog/code-navigation-give-your-agent-real-access-to-the-source/): A skill that gives agents search, file listing, read, and grep across a codebase.
- [Package Intelligence](https://githits.com/blog/package-intelligence-vet-a-dependency-before-you-adopt-it/): License, dependency, vulnerability, changelog, and upgrade-review data before you add a dependency.
- [Compare Coding Agents by How They Actually Behave](https://githits.com/blog/compare-coding-agents-by-how-they-actually-behave/): Use GitHits in the research phase to compare agents by real behavior before cloning a repo.
- [Context is King #3: Open-Source Code as Context](https://githits.com/blog/open-source-code-context-for-coding-agents/): Olli-Pekka Heinisuo on how open-source code becomes useful context for coding agents.
- [Context is King #4: Source-Grounded Context](https://githits.com/blog/source-grounded-context-for-coding-agents/): Nathan Burg on source-grounded context and why real data matters for trustworthy AI workflows.

## Case studies

- [DuckDB API Migration](https://githits.com/blog/duckdb-api-migration-case-study/): How GitHits surfaced the DuckDB v1.3.2 source behavior a vanilla agent run missed.
- [SQLx 0.9 Safety Migration](https://githits.com/blog/sqlx-querybuilder-safety-case-study/): How GitHits preserved SQL injection safety while cutting search churn in an SQLx 0.9 migration.
- [Pydantic AI Tenant Routing](https://githits.com/blog/pydantic-ai-tenant-routing-case-study/): How GitHits grounded a Pydantic AI tool fix in current docs and source while cutting token churn.

## Community

- [Wall of Love](https://githits.com/wall-of-love/): Real reviews from developers using GitHits, sourced from X, Reddit, LinkedIn, GitHub, Hacker News, and personal blogs.

## Optional

- [Terms of Service](https://githits.com/legal/terms-of-service/)
- [Privacy Policy](https://githits.com/legal/privacy-policy/)
- [Cookie Policy](https://githits.com/legal/cookie-policy/)
- [Data Processing Agreement](https://githits.com/legal/data-processing-agreement/)
- [Subprocessor List](https://githits.com/legal/subprocessor-list/)
- [RSS feed](https://githits.com/rss.xml): Blog updates.
- [Sitemap](https://githits.com/sitemap-index.xml): Full list of indexable pages.