GitPulse
AI-powered Git guardrails that learn your team's conventions from commit history. An MCP server with 10+ tools, a React Ink CLI, and a Next.js analytics dashboard — all working together to enforce quality without slowing you down.
By the Numbers
- 18K+ — Lines of Code
- 89 — Commits
- 10+ — MCP Tools
- 3 — Surfaces (CLI, Web, MCP)
The Problem
Code review catches convention violations after the code is written. Linters catch syntax, not intent. Teams waste cycles arguing about commit message formats, branch naming, and PR structure — none of which existing tools handle intelligently.
GitPulse learns your team's actual conventions from commit history and enforces them at the point of creation, not after the fact.
Architecture
Three interfaces, one engine. The MCP server is the core — it exposes 10+ tools that any AI agent can call to analyze commits, suggest improvements, and validate conventions. The React Ink CLI gives developers a terminal-native experience. The Next.js dashboard provides team-wide visibility into convention adherence and code quality trends.
The convention learning engine analyzes historical commits to build a model of what "good" looks like for your specific codebase — not generic rules, but your team's actual patterns.
Key Features
- Convention Learning — Analyzes commit history to learn team patterns — message formats, branch naming, file organization. Adapts to your workflow, not the other way around.
- MCP Tool Suite — 10+ tools exposing commit analysis, convention checking, quality scoring, and suggestion generation to any MCP-compatible AI client.
- React Ink CLI — Beautiful terminal interface built with React Ink. Interactive commit message suggestions, branch name validation, and pre-commit checks.
- Next.js Dashboard — Web dashboard for team-wide analytics. Convention adherence trends, quality metrics, and contribution insights.
Tech Stack
TypeScript · Next.js · React Ink · MCP SDK · Node.js · Tailwind CSS · Netlify