Open data & API
The entire dataset behind this site is public: a versioned static JSON API, Atom feeds, embeddable badges, and llms.txt. No key, no rate-limit games, no tracking — and because there is no paid placement, what you pull is what the evidence supports.
Last reviewed · dataset license: MIT (each record's own software license is a field on the record) · schema: data/SCHEMA.md
JSON API (static, versioned)
Every endpoint is a build-time artifact — plain JSON on a CDN, cacheable forever, fetchable from browsers, CI jobs, and agents alike. Start at the manifest.
| URL | What it is |
|---|---|
| /api/v1/index.json | Manifest: schema version, counts, every endpoint, and all tool ids |
| /api/v1/tools.json | All tracked tools with full source-backed governance posture |
| /api/v1/platforms.json | The cloud foundation platforms |
| /api/v1/updates.json | The full source-backed change log |
| /api/v1/tools/crewai.json | Any single tool by id (see index.json for the id list) |
| /api/v1/badge/crewai/license.json | shields.io endpoint badge — also status.json and verified.json |
Feeds & calendars
| URL | What it is |
|---|---|
| /updates.xml | Site-wide updates feed (Atom) |
| /updates-agents.xml | Agents only — also orchestration, governance, assistants, platforms |
| /updates-licenses.xml | License changes only — the relicensing alert channel |
| /eu-ai-act-deadlines.ics | EU AI Act deadline calendar (iCalendar) |
| /llms.txt | Index for LLM agents; llms-full.txt is the whole site in one fetch |
Embeddable badges
Live, source-backed badges for any tracked tool via the shields.io endpoint schema: license, status, and verified (the dated governance-review stamp). Drop this in a README:
Swap crewai for any tool id from the manifest, and license for status or verified.
Stability & versioning
- Endpoints under /api/v1/ are additive: fields may be added, existing fields are not renamed or removed within v1.
- Record ids are permanent. A renamed product keeps its id; prior names are preserved as aliases.
- Every artifact is regenerated from the canonical dataset on each build and is deterministic — identical data produces identical files.
- Breaking changes would ship as /api/v2/ with v1 kept alive; there are currently no plans for a v2.
Consuming the data somewhere interesting? Open an issue in the repository — knowing what depends on v1 is how we keep the stability promise.