How we curate AI Tools Nav (sources, caps, attribution)

AI Tools Nav is a curated English directory, not a scrape of every SaaS that added “AI” to its landing page. Here is how the seed catalog is built and refreshed.

Sources

• OSS AI Hub (ossaihub.com) — open tools JSON, CC BY 4.0. Attribute when you reuse.
• aifoxx (MIT) — broader AI product list we map into our categories.
• Editorial keep/drop lists in scripts/import-oss-seed.mjs so infra noise and non-AI tools stay out.

Caps (why the directory stays browsable)

• Target 80–120 tools in the local seed snapshot (src/data/seed/items.json).
• Per-category caps so one vertical cannot swallow the homepage.
• A fixed featured set for the homepage (ChatGPT, Claude, Gemini, Midjourney, Cursor, Perplexity, Hugging Face, Ollama, Copilot, Runway, ElevenLabs, Gamma).

Refresh path

pnpm seed:import

Or run the scheduled GitHub Action Sync OSS seed catalog (weekly + manual). Review the PR if the workflow opens one; merge when the diff looks sane.

When Sanity is configured, pnpm batch:import can push structure into the CMS — that path needs SANITY_API_TOKEN and is separate from the seed JSON.

What we optimize for

• Job-to-be-done categories (writing, image, code/agents, research…) over alphabet dumps
• Honest pricing tags (free / freemium / paid / open-source)
• Short descriptions you can scan in under a second

What we refuse

• Paying for rankings without disclosure
• Inflating counts with browser extensions and “AI” CRMs
• Copying competitor blurbs verbatim

Homepage: aitoolsnav.com. Categories start at /category.