Introduction to GitVelocity
What Is GitVelocity?
GitVelocity is an AI-powered engineering velocity measurement platform. It scores every merged pull request on a 0--100 scale using Claude, giving engineering teams an objective, consistent way to understand the complexity of the work they ship.
Unlike estimation-based approaches that try to predict effort before work begins, GitVelocity measures complexity after code ships. The result is a ground-truth record of what actually happened -- not what someone guessed would happen.
Measure what ships.
How Scoring Works
When a pull request merges, GitVelocity analyzes the diff across six dimensions:
| Dimension | What It Captures |
|---|---|
| Scope | Number of files, lines changed, and breadth of the change |
| Architecture | Structural impact -- new modules, API changes, cross-boundary modifications |
| Implementation | Algorithmic complexity, novel logic, and depth of reasoning required |
| Risk | Potential for regressions, concurrency issues, and data integrity concerns |
| Quality | Test coverage, documentation, error handling, and observability |
| Performance / Security | Optimization work, caching strategies, authentication, and input validation |
Each dimension receives a sub-score, and the final score is a simple sum -- no hidden weights, no opaque formulas. A routine config change might score in the single digits. A complex distributed-systems feature might land in the 80s or 90s.
Who Benefits
Engineering Leadership
Engineering managers and directors gain objective visibility into what their teams deliver. GitVelocity replaces anecdotal impressions with data, making it easier to have informed conversations about workload distribution, capacity planning, and team performance over time.
Individual Engineers
Engineers can demonstrate the impact of their contributions with concrete metrics. Difficult, high-complexity work becomes visible regardless of whether it results in a flashy feature or a critical behind-the-scenes improvement.
Organizations
Teams and organizations develop a shared understanding of engineering output. When everyone references the same scoring framework, discussions about priorities, trade-offs, and resource allocation become more productive.
Core Principles
Transparent scoring. The six dimensions and the sum formula are fully visible. There is no black box.
Post-merge measurement. GitVelocity scores what was actually built, not what was planned. This removes the biases and inaccuracies inherent in upfront estimation.
AI-powered consistency. Claude applies the same analytical framework to every PR, eliminating the variability of human reviewers estimating complexity differently.
Privacy-first. GitVelocity reads PR diffs for analysis but does not store your source code. See Connecting Your Repositories for details on permissions and data handling.