GitVelocity

About GitVelocity

GitVelocity scores merged pull requests with an LLM and turns the result into a single engineering-output number, positioned explicitly as a replacement for story points in a world where AI assistance has broken the relationship between effort and line count. You connect GitHub repositories and every merged PR is analysed across six dimensions of complexity, producing a score that reflects what the change actually did rather than how many lines it touched or how long a ticket sat in a column. The company published an analysis of 50,000 scored PRs as its launch artefact, which is the sort of evidence that is rare in the developer-productivity category. There are two audiences and two views: individual engineers track their own velocity over time and can see how each contribution was scored, which makes the metric arguable rather than imposed; engineering leaders compare output across teams and time periods without needing to sit in standups to know what is moving. The measurement stance is deliberate — the marketing rejects vanity metrics and estimates in favour of scoring only what actually merged. The commercial model is the unusual part for this category: the product is free with no paid tier, and you supply your own LLM API key, so the marginal cost of analysis lands on your provider bill rather than on a per-seat subscription. That makes it viable for a three-person team and for a three-hundred-person org on the same terms.

Key Features

GitVelocity Pros & Cons

Tags