Pricing

One verification = 5–8 AI calls across 3 independent models from different providers.

We pay for AI calls. You choose the models. No API keys needed.

Bring your own API keys. Pay providers directly. Unlimited verifications.

Free
See where AI gets it wrong.
$0
3 managed + 5 BYOK verifications
No credit card required
20 Quick + 5 Standard per month
No credit card required
Expert
Where one mistake costs thousands.
$149/mo
~300 verifications with mid-range models
Full verification pipeline
$99/mo
Unlimited verifications
Full verification pipeline

🔬 Founding Member — first 100 users

Free BYOK access during closed beta 20% off any paid plan for 12 months Direct access to founders for feedback and feature requests
73 of 100 spots left

Team

Custom pricing · From $99/seat · BYOK or Managed

Everything in Expert + Shared workspace + Audit logs + Dedicated onboarding Get a quote →

Frequently asked questions

Everything you need to know about CrossCheck pricing.

BYOK vs Managed
What's the difference between BYOK and Managed?
BYOK (Bring Your Own Key): You connect your own API keys from Anthropic, OpenAI, and Google. You pay AI providers directly for the tokens you use. CrossCheck charges only the platform subscription — verifications are unlimited.

Managed: We handle everything. No API keys, no setup. We pay for AI calls on your behalf. Your plan includes a monthly usage budget — how far it stretches depends on which models you choose.

In both modes, you have full control over which specific model to use from each provider.
Can I switch between BYOK and Managed?
Yes, at any time. It's a plan change — pricing adjusts at the next billing cycle. Many users start with Managed for simplicity, then switch to BYOK once they're comfortable with API keys and want unlimited verifications at a lower price.
Usage & Models
How does Managed usage work?
Each Managed plan includes a monthly API budget. You choose your models from each provider — and that choice determines how far the budget goes.
Premium models (Opus, Sonnet) — most capable, use budget faster. ~35–50 verifications on Pro.

Mid-range models (GPT-5.4, Gemini Pro, DeepSeek R1) — great quality-to-cost balance. ~100–130 verifications on Pro.

Economy models (Haiku, Flash, Flash-Lite) — fast and efficient. ~500+ verifications on Pro.
You can mix and match across sessions. Track your remaining budget in real time in the dashboard. We notify you at 80% usage — no surprise charges.
How much do API keys cost in BYOK mode?
Depends entirely on the models you choose. A typical Standard verification (3 models + adversarial rounds):
Economy mix (Haiku + Flash-Lite + Flash): ~$0.01–0.04 per session

Mid-range mix (GPT-5.4 + DeepSeek + Gemini Pro): ~$0.15–0.25 per session

Premium mix (Opus + Sonnet + GPT-5.4): ~$0.30–0.65 per session
At 50 sessions/month with mid-range models, expect ~$8–12 in total API costs paid directly to providers. Longer documents use more tokens and cost more per session. We show estimated cost before each verification starts.
What are the verification presets?
All presets use 3 models from different providers. You always choose which specific model from each.
Quick — Fast parallel comparison. Three models answer independently. Best for simple fact-checks.

Standard — Three models answer, then challenge each other through adversarial rounds. Produces a disagreement map showing where and why models conflict.

Deep — Standard + ChainPoll: each individual claim in the final answer is verified separately 3 times by an independent model. Flagged claims get a reliability score.

Deep+ — Full multi-agent verification pipeline with additional specialized agents. Expert tier only.
Billing
What happens if I use up my Managed budget?
We notify you at 80% usage. When the budget is fully used:
1. Switch to BYOK — unlimited verifications at a lower subscription price.

2. Upgrade your tier — Expert includes 3× the budget of Pro.

3. Wait for next cycle — your budget resets at the next billing date.
We never charge overages without your explicit consent.
Do long documents cost more than short questions?
Yes. AI models charge by tokens (roughly 750 words ≈ 1,000 tokens). A short question uses ~2,000 tokens per model call. A 20-page document uses ~25,000.

In Managed mode, longer documents use your budget faster. In BYOK mode, your provider bill reflects actual token usage. Either way, we show the estimated cost before you start each verification — no surprises.
What does the Founding Member offer include?
The first 100 users who join the beta get three benefits:
Free BYOK access — use CrossCheck with your own API keys at no charge during the entire beta period.

20% off for 12 months — when you move to a paid plan, get 20% off any tier (Pro $49→$39, Expert $149→$119) locked for a full year.

Direct founder access — private feedback channel with the founding team. Your input shapes what we build next.

Join the first 100

Free BYOK access during beta. 20% off for 12 months. Direct access to founders.

Join the waitlist →

Enterprise and Academic plans available. [email protected]