Pricing
The detector is free to use today — no credit card required. Priced tiers with larger quotas will be available soon.
First release targets English-language hallucinations. Code and other domains follow.
3,700 API detections/month · 300 web-app detections/month · up to 2,048 characters per detection · no credit card
Detector (web) — paid plans
Available soonFor people checking LLM output in a browser.
Pro
For regular users
Writers, editors, fact-checkers using it daily
- 6,000 detections/month (~200/day average)
- Up to 2,048 characters per detection
- Hard cap — no overage charges
- Priority web rate limits
Premium
For heavier daily use
Power users running large batches
- 12,000 detections/month (~400/day average)
- Up to 2,048 characters per detection
- Hard cap — no overage charges
- Priority web rate limits
API (for developers) — paid plans
Available soonFor integrating hallucination detection into your own product. Token-style usage metering; published rate limits; monthly quota with pay-as-you-go overage.
Starter
Small apps
Indie projects, early-stage products
- 10,000 detections/month
- Up to 2,048 characters per detection
- Overage: $6 / 1K detections
Growth
Established apps
Production teams with steady volume
- 75,000 detections/month
- Up to 2,048 characters per detection
- Overage: $4 / 1K detections
Scale
High volume
Heavy API traffic
- 360,000 detections/month
- Up to 2,048 characters per detection
- Overage: $3 / 1K detections
What counts as a detection?
1 detection = analyzing one LLM response of up to 2,048 characters. The character limit counts the response plus the original prompt, if provided. This applies to both the web detector and the API.
How API billing works
Web tiers (Web Pro / Web Premium) use daily caps with hard cut-offs, not monthly overage. The Free tier splits its allowance between the two surfaces (3,700/mo via API, 300/mo via the web app — separate caps). The bullets below describe paid API tier behavior only.
- Per-call billing header. Each
/detectresponse includes anX-Detections-Billedheader so you always see what you were charged. - Overage, not cliffs. Go past your monthly quota and you pay the per-1K-detection overage rate listed on your tier — no cut-offs.
- Rate limits are published. Every API response carries
X-RateLimit-*andX-Quota-Remainingheaders so your client can back off gracefully.
Have questions? See the FAQ →