API pricing
Every request returns facts graded on how distinct their sources are — not just source count. We group sources by website and by parent company for the websites we have mapped, set aside Wikipedia mirrors when sources of another kind exist, and rank each claim Gold / Silver / Bronze on the resulting source groups.Read the verification methodology →
Free
Standard-quality facts, community support, attribution required
- ✓500 requests / month
- ✓30 requests / minute
- ✓Silver and Gold tiers
- ✓Cited sources on every fact
- ✓Confidence score & category tags
Starter
Standard + verified facts, 60 req/min, email support, commercial use OK
- ✓10,000 requests / month
- ✓60 requests / minute
- ✓Silver and Gold tiers
- ✓Cited sources on every fact
- ✓Confidence score & category tags
- ✓Cancel anytime, prorated
Growth
All fact tiers + audit mode, 300 req/min, priority email support
- ✓50,000 requests / month
- ✓300 requests / minute
- ✓Gold + Silver + Bronze verified facts
- ✓Cited sources on every fact
- ✓Confidence score & category tags
- ✓Cancel anytime, prorated
Scale
All facts + bulk export + webhook alerts, 1000 req/min, shared Slack channel
- ✓250,000 requests / month
- ✓1000 requests / minute
- ✓Gold + Silver + Bronze + audit mode
- ✓Cited sources on every fact
- ✓Confidence score & category tags
- ✓Cancel anytime, prorated
Enterprise
Custom quota, SLA guarantees, dedicated infrastructure, SSO/SAML, MSA/DPA — contact sales
- ✓Custom monthly quota
- ✓Custom rate limits
- ✓Gold + Silver + Bronze + audit mode
- ✓Signed uptime SLA
- ✓Private endpoints
- ✓Dedicated Slack support
Frequently asked
How is a "verified fact" different from a normal AI response?
Every fact is graded on how distinct its sources are — not just corroboration count. Our pipeline groups sources by website, and by parent company for the websites we have mapped, sets aside Wikipedia mirrors and aggregators when sources of another kind exist, and computes a 0–1 score from how the remaining groups differ in website or mapped parent company and in kind of source, weighted by kind of source (the field is named independence_score; the name is historical). Facts are then labeled Gold, Silver, Bronze, or Unverified. Free and Starter plans return Silver and above only. Growth adds Bronze for content platforms doing broad discovery. Scale and Enterprise add opt-in audit mode with flagged unverified claims. See the verification methodology.
Can I cancel any time?
Yes. Manage your subscription from the dashboard — cancel, upgrade, downgrade, or update your card. No forced yearly plans.
What happens if I exceed my monthly quota?
You'll get a 429 response with a Retry-After header telling you when the quota resets. Upgrade in-flight from your dashboard to keep serving traffic — the new quota is available immediately.
What are the rate limits for?
The per-minute limits protect the service from runaway loops, not you. If you hit them accidentally, wait 60 seconds and retry — no penalty. Sustained traffic that needs a higher ceiling belongs on Growth (300 req/min) or Scale (1,000 req/min).
Do you offer a free trial?
The Free plan is our trial — 500 requests every month, permanently free. Ship a real integration on it, and upgrade only when your traffic outgrows the tier.
What does Enterprise include?
Custom monthly quota, private endpoints, a signed uptime SLA, dedicated Slack support, and audit-mode access to flagged unverified claims. Get in touch at api@deepinquiry.ai and we'll size a plan to your traffic.
What is Gold vs Silver vs Bronze?
Gold = at least three separate source groups (four for exceptional claims), a high score, at least one primary or peer-reviewed source, and no Wikipedia, mirror or aggregator source removed. Some claim types are capped at a lower tier. Silver = at least two separate source groups at the standard bar (three for exceptional claims). Bronze = at least two groups with a lower score. Tiers measure how many different websites and owners stand behind a fact, not whether the evidence is independent. Unverified = below Bronze; never returned by default. Read the full spec in the methodology page.