Company API guide

Company API plans

Both tiers return the same stored company profiles and catalyst summaries through GET /api/v2/companies. One delivered company costs one credit per export, across all its pages.

All keys on an account share its allowance and request limits. There are no automatic overages. Errors, estimates, page retries and unchanged conditional requests use no company credits, but count toward request limits.

The billing adapter accepts Stripe test mode only. Live purchases are unavailable.

Company API access, credits and response guide · Public v1 preview reference

Catalyst summaries

Both v1 and v2 return these fields on each item in companies[].catalysts[]. The keys are always present; their values can be null. A summary describes one catalyst, such as an FDA decision. There is no separate company-level AI overview in either response.

summary

A stored plain-English admin or research note takes priority when it passes the public output rules: at most 400 characters and no URL. Otherwise the API builds a deterministic explanation from the published catalyst fields. If those fields cannot support an explanation, the value is null.

summaryBy

admin means an editor-written stored note. research means a stored research-pipeline note, including AI-generated explanations written earlier. template means a deterministic explanation from catalog fields, not an AI response. Null means no summary was published.

summaryAt

The stored note’s write time, when available. It is null for templates and seed sentences without a reliable authorship timestamp. It is not the FDA decision date, source verification time or API response time.

API reads never call a model or trigger research. An AI summary is available only if it was already written and stored; it is not guaranteed for every catalyst. A template fallback may be assembled during serialization without an AI call. V2 uses its immutable snapshot’s publication time when wording that fallback.

Use source, sourceUrl, sourceClass, verified, verifiedAt and confidence for the separate source and verification context. Summary attribution does not prove the underlying event was independently checked. Provider names and model IDs are not exposed per summary. V2’s optional profileFacts contains selected stored facts, not a company AI overview.

Fictional FDA decision example

This is a serializer-produced example for fictional ticker EXMP and fictional treatment Examplemab. The text, dates and example.test link illustrate a stored research summary; they are not real FDA data, a model run or a prediction. It shows one catalyst object, not a complete response.

{
  "id": "exmp000000000001",
  "type": "pdufa",
  "title": "Fictional catalog label",
  "drug": "Examplemab",
  "indication": "Example condition",
  "phase": "Phase 3",
  "date": "2026-11-14",
  "datePrecision": "day",
  "dateStart": "2026-11-14",
  "dateEnd": "2026-11-14",
  "nctId": null,
  "outcome": null,
  "confidence": null,
  "summary": "In this fictional example, the FDA decision for Examplemab is scheduled for November 14, 2026. This stored explanation does not predict the outcome.",
  "summaryBy": "research",
  "summaryAt": "2026-09-28T12:00:00.000Z",
  "source": "curated",
  "sourceUrl": "https://example.test/fda/decision",
  "sourceClass": "company_or_other",
  "verified": false,
  "verifiedAt": null
}

If no eligible stored note exists, the same catalyst can instead have summaryBy: "template" and summaryAt: null. If neither a note nor a valid fallback is available, all three summary fields are null.