KodHau
MCP tool that feeds tribal knowledge to AI codersGive AI agents the context your PRs already contain
Highlights
What you can do with it
- Teams stop AI agents from repeating rejected approaches
- Individual developers mine 30 recent PRs for historical context
- Enterprises keep code private with SSO and private deployment
How to use it
- 1 Let KodHau scan your closed PRs for decisions and constraints
- 2 Enable the MCP integration so context is injected before code is written
- 3 Choose an MCP-compatible client such as Cursor or Claude Code
- 4 Pick the free, Pro, Team, or Business tier that fits your team
Who it is for
Pros
- Free version for individual developers to explore
- Local execution keeps code and PR data on your machine
- Reduces AI mistakes caused by missing background
- Fits into existing development workflow
Limitations
- The free version only covers 1 repository and 30 PRs; paid tiers cost money and free rules can change
- It depends on your PR history, so thin or messy histories give thin context
- Needs an MCP-compatible client and some setup
- Local runs mean you maintain the tool and its updates
What KodHau MCP is
KodHau MCP is a tool that gives AI agents the tribal knowledge your team already has. It scans closed pull requests and pulls out architectural decisions, code review comments, rejected approaches, and constraints. Through the Model Context Protocol, it injects that context into an AI agent before the first line of code is written, so the agent understands why things are the way they are. The site says it runs locally and that code and PR data do not leave your machine. There is a free version for individual developers plus paid Pro, Team, and Business tiers, and it plugs into clients such as Cursor and Claude Code.
What you can do with it
- Analyze PR history to recover decisions, review comments, and rejected methods
- Inject constraints and architectural context into AI agents before coding
- Work inside MCP-compatible clients such as Cursor and Claude Code
- Start free with one repository and the latest 30 PRs, then scale up to team and enterprise tiers
Who it is for
- Individual developers who want project history inside their AI prompts
- Teams that need shared tribal knowledge so new members ramp faster
- Enterprises that require SSO and private deployment for code and data control
What to watch out for
The free version is limited to one repository and the latest 30 PRs, so larger projects need a paid tier, and free terms can change. The quality of the injected context depends entirely on your PR history: if reviews and decisions are thin or inconsistent, the AI gets thin, inconsistent guidance. It requires an MCP-compatible client and some setup, and because it runs locally you are responsible for keeping it updated. Treat it as a context aid, not a replacement for reading the code yourself.
FAQ
What does KodHau MCP do?
It extracts tribal knowledge from PR history and injects it into AI agents before they write code.
Does my code leave my machine?
The site says it runs locally and code and PR data do not leave your machine.
What does the free version include?
Individual developers can analyze 1 repository and the latest 30 PRs.
Which clients are supported?
Any client supporting the Model Context Protocol, such as Cursor and Claude Code.
Why it is free
What the official page says： free version
Free limits, trial length and features are decided by the official site and can change at any time.
Product preview
Screenshot archived from the listing page, for preview only.
