InfoReader is agent-readable
Everything on this site is also published as open, static JSON. No key, no account, no scraping needed — the same feeds the site itself renders.
Feeds
| Full briefing feed | /data/latest.json | generated_at, stats, briefing, by_topic |
| Top briefing only | /data/briefing.json | The ~12 lead stories |
| Per-topic desks | /data/topics/<topic>.json | world, business, ai, science, health, general |
| Single story | /data/stories/<id>.json | Full record incl. every source link |
| API description | /openapi.json | OpenAPI 3 spec for the feeds above |
| Plain-text guide | /agents.md • /llms.txt • /llms-full.txt | This guide in Markdown / LLM formats |
Story records
Each story includes its headline, summary, key points, how the summary was made (summary_method: extractive = composed from the sources' own text; groq / gemini / workers_ai = that AI tier), a source count, and the full list of original articles. For the underlying reporting, cite the original publishers — they are linked in every record.
Honest limits
The API is read-only: no posting, no accounts, no personalization. Email signup happens through the web form on this site (backed by Brevo from Phase 3) — there is no subscription API yet. Feeds refresh whenever the pipeline runs, roughly hourly in production; generated_at is your freshness check. The feeds are static CDN files with no published rate limit — please cache and poll politely.