Run the GitDealFlow Momentum Score on any public GitHub organisation or repository, live from the GitHub API. Commit acceleration, contributor growth, bus factor and repository expansion — the engineering signals that move before the round, not after it.
api.github.com directly. Nothing is routed through GitDealFlow, nothing is stored, and no token is requested — only public repository data is read. Check it in your network panel.GitDealFlow's premise is that engineering activity is a leading indicator: a team ships harder before it raises, and that shows up in public commit history weeks before a press release. This tool runs that read on demand for any public org or repo, using the same score as the public Momentum Index so results are directly comparable.
Identical to the one the Momentum Index publishes:
traction = 100 × log10(stars + 1) / 5 — 100,000 stars maps to 100.
recency — piecewise on days since the last push: ≤2 days → 100, ≤7 → 88, ≤30 → 68, ≤90 → 40, ≤180 → 22, otherwise 8.
SCORE = round(0.40 × traction + 0.35 × recency + 0.25 × velocity)
The index computes velocity from week-over-week star growth against its own stored snapshots. A one-off check of an arbitrary repository has no such history, so this tool substitutes commit acceleration — derived from GitHub's 52-week commit series, which needs no history at all — and states on the result which velocity term was used. Where the index would fall back to traction, the substitute is strictly more informative: star growth is an audience signal, commit acceleration is an engineering one, and the site's whole thesis is that the engineering signal moves first.
acceleration = (mean weekly commits, last 4 weeks) ÷ (mean weekly commits, the 12 weeks before that), from /stats/commit_activity.
Volume alone rewards large mature projects that are not changing pace. Acceleration isolates the change in tempo. Above ~1.3 the team has genuinely stepped up; below ~0.7 output is decaying, which on a well-known repository is often the more interesting finding.
From /stats/contributors (top 100 contributors, weekly history). Active contributors are those with at least one commit in the last 8 weeks. New contributors are those whose first-ever commit week falls in the last 12 weeks — a team that is hiring shows up here before it shows up on LinkedIn. Bus factor is the share of the last 52 weeks' commits authored by the single most active contributor; above ~60% the project is one person, which is a diligence question no matter how fast it moves.
For an organisation, the three most-starred repositories pushed in the last 180 days are scored — ranking purely by push recency picks whatever happened to be touched last, which is often a docs or config repo rather than the company's actual product. The org's public repo list is also summarised: how many were created in the last twelve months and how many were pushed in the last thirty days. A company spinning up new public repos is usually building surface area — new SDKs, integrations, docs sites — which tends to accompany a commercial push.
Only three repositories are fetched per organisation check, to stay inside GitHub's unauthenticated budget of 60 requests per hour. The count is stated in the result rather than silently truncated.
Public activity only: private monorepos, GitLab, and companies that develop in the open only for a thin SDK will all read low. Contributor statistics cover the top 100 contributors. GitHub's /stats/* endpoints compute asynchronously and occasionally return "not ready" on a first call for a very large repository — retry. And the obvious one: many excellent companies have almost no public code, so a low score is evidence of absence of a public signal, not evidence of absence of momentum.
The public GitHub REST API, called directly from your browser. Nothing is proxied through GitDealFlow and no result is stored. Only public repository metadata is read — stars, last push date, weekly commit counts and contributor statistics. Private repositories are not accessible and no authentication is used.
GitHub allows 60 unauthenticated API requests per hour per IP, and each check spends three to six. The remaining allowance is shown after every check and the limit resets hourly. This tool deliberately does not ask for a personal access token — pasting credentials into a third-party page is a habit worth not building.
No. The Momentum Score reflects public software-engineering activity only. It is not investment advice, a valuation, or a recommendation to buy, sell, or fund any company. It is one leading indicator and says nothing about revenue, retention, market or team.
This checks one name at a time, when you think to check it. The weekly signal watches 4,200+ organisations continuously and tells you which ones changed — the difference between looking something up and being told where to look.
The public Momentum Index · All free tools · Startup calculators · Startup signal checklist · VC deal-flow tracker