Developer Tools
Graft
Last updated 2026-09-21 · benchmark measured 2026-11-03 — deterministic & reproducible
Graft is an open-source context layer for large codebases used with AI coding agents.
Graft at a glance
Graft is an open-source tool that maps the structure of large codebases so AI coding agents can understand and navigate the code, built for developers, engineering teams, and codebase maintainers.
Good for
- Helping AI coding agents like Claude Code, Codex, Cursor, or Kiro understand a large codebase without reading every file
- Checking the impact of a code change before editing, by seeing what a symbol depends on and what depends on it
- Keeping a codebase's map up to date automatically during coding sessions, without manually re-running anything
Pricing
Free (open source, MIT license).
Questions people ask
Is Graft free to use?
Yes, Graft is free and open source under the MIT license.
Does Graft send my code to a server or the cloud?
No, Graft runs 100% locally with no vector embeddings and no telemetry; it works with plain files your agent reads.
How do I set up Graft?
You install it with 'npm install -g @nanonets/graft' and run 'graft init' once; after that it rebuilds the map in the background during every session automatically.
Which AI coding agents does Graft work with?
It works with Claude Code, Codex, Cursor, Kiro, and any other agent that reads files, using an MCP server and CLI with six MCP tools.
What performance difference does Graft make?
According to benchmarks across 162 runs on two repos, using Graft's map instead of a cold start reduced latency by 60%, tool calls by 46%, input tokens by 42%, and cost per task by 32%.
Written by Legit.Show from trailhq.com on 2026-09-30. It does not change the score.
Are you the maker of Graft? Claim this listing. It is free, takes a meta tag, a DNS record or GitHub admin rights, and never changes the score.
Launched something? Add your product, free.
To cite this score: legit.show/s/gh-trailhq-graft/2026-11-03. That address never changes; this page moves with every re-measure.
Legit.Show scored Graft 83/100 on 2026-11-03, measured across all 7 frames from its public surface. legit.show/s/gh-trailhq-graft/2026-11-03
Is Graft production-ready?
Legit.Show scores Graft 83 out of 100 — the simple average of its 7 measured frames. Legit.Show ran its deterministic 7-Frame production-readiness benchmark on Graft (public-surface assessment), measured from the public surface with no LLM in the scoring path. Its strongest frame is Accessibility; its weakest is Security. Every frame it averages is published with its evidence on the Legit.Show listing.
The 7 Frames
- Performance — 75/100
- Accessibility — 100/100
- Security — 45/100How this compares across the catalogue: The Web Security Baseline
- Privacy — 70/100How this compares across the catalogue: The Privacy Gap
- Reliability — 100/100
- Standards — 92/100
- Discoverability — 100/100
What we measured
- Security headers present: HSTS.
- No Content-Security-Policy.
- Served over HTTPS with a valid certificate.
- Real Lighthouse performance run — 188 ms to first byte.
- Returns a proper 404 for unknown routes.
- 3 of 3 sampled routes reachable.
- Has a reachable privacy policy.
- Sets cookies / loads scripts with no consent prompt.
Who it's for
Developers · Engineering teams · AI agent users · Codebase maintainers
Pricing
Free (open source)
Sources and updates
- Description
- Taken from trailhq.com's own website on 2026-09-21.
- Operator
- Trail, Inc., as named in trailhq.com's terms, privacy page or footer.
- Benchmark
- Measured by Legit.Show from the live site, the way any visitor sees it — with no access to its code or accounts. Same method for every product, and no AI decides the score. Last checked 2026-11-03.
Put together from public information, without Graft's involvement. If anything here is wrong, tell us and a person will check it.
Visit Graft → · Alternatives to Graft → · How this was measured →
