Verdict-first
A fit call — strong / solid / care / risky — plus a one-line bottom line, before anything else. The decision, then the evidence.
Case No. RL-3.0 · Chrome · Manifest V3Read code before
Read code before
you trust it.
One click turns any GitHub, GitLab, npm, or PyPI page into a plain-English, verdict-first briefing — what it is, whether it fits, and how it’s actually built. Not the README’s pitch. The real shape of the thing.
Bring your own model · 20+ providers · no server, no account, nothing leaves your browser
Lenses up — reading past the README.
The verdict
It opens with the verdict, not the pitch.
Stars tell you a project is popular. They don’t tell you whether it fits your problem. RepoLens leads with a fit call and a one-line bottom line — then lets you go as deep as you want.
- A fit call first — strong, solid, care, or risky — before a word of prose.
- Measured, not trusted — health, languages, flags, and tech stack pulled from the source.
- Deep on demand — ELI5, lineage, alternatives, and a Feynman-style explanation one click away.
expressjs/express
Fast, unopinionated, minimalist web framework for Node.js.
JavaScript65k ★MIT
88
Repo health
Strong fit
AI bottom line
The default choice for Node HTTP services; boring in the best way.
JavaScript 99%TypeScript · Shell
The case file
A research tool, not a bookmark folder.
Thirty-plus releases of compounding workflow — from the first verdict to a library that finds its own peers.
Discovery & recommendations
Search GitHub from inside the extension, or let RepoLens recommend peers from the repos you’ve already adopted — same capabilities, same language, ones you haven’t seen yet.
Library + triage
Every scan becomes a sortable, filterable grid. Decide with the keyboard, in flow.
Boards
Group the repos you’re weighing into collections you can compare and revisit.
Evaluations
Score repos 1–5 against your own weighted rubric. The badge follows each card.
N-way compare
Put any 2–10 repos side-by-side in a structured matrix; export to CSV or Markdown.
Deep Dive
Atoms → lineage → a Feynman-style explanation, optionally grounded in measured facts from the source: real file counts, the dependency graph, tests, and a secret scan.
Drift alerts
A daily background check flags repos that have gone stale, so your shortlist stays honest.
Bring any model
Your keys, 20+ providers, route each part of a scan to a different model — or run local Ollama for $0.
The procedure
Three seconds from page to verdict.
- 01
Land on any repo
GitHub, GitLab, an npm package, a PyPI project. Wherever you’re evaluating, RepoLens is one icon away.
- 02
Click once
It reads the source and runs it past the AI provider of your choice. The optional runner grounds it in measured facts — never executes repo code.
- 03
Get the verdict
A fit call, health, deep dive, and red flags — opened in a tab and saved to your local library to sort, compare, and revisit.
Models & privacy
Bring your own model. Keep your own data.
Route the verdict to a fast model and the deep dive to a strong one, with smart fallback. Twenty-plus providers behind one registry — any OpenAI- or Anthropic-compatible endpoint just works.
ClaudeGPTGeminiGrokDeepSeekGroqQwenKimiGLMNVIDIA NIMOllamaCustom
- No backendThere’s no server and no telemetry. The whole library lives in your browser.
- Your keysKeys are stored locally and sent straight to the provider you chose — never to us.
- $0 pathRun a local Ollama model or a free Gemini tier and pay nothing at all.
The detective
Vee reads every case.
The lens mascot reacts to what it finds — wide-open on a clean repo, narrowed and skeptical on a risky one. Every expression maps to a real scan moment, and all of it folds to a static glyph under reduced-motion. Works the day shift or the night stakeout.
On the case
Reading
Deep dive
Cleared
Suspicious
Off duty
Stop trusting the README.
Install RepoLens, point it at the next dependency you’re weighing, and read the real shape of the thing — in plain English, in seconds.