Pricing that scales with you.
Choose your guardrail. Free for individuals, paid tiers for teams that ship to production.
Free
Initium
Free to use. No card required.
- Token estimates on all models
- Basic cost forecasting
- No account required
Pro
Certitudo
For developers shipping AI features.
- Everything in Free
- At least 20x more analyses than free - even during peak demand
- Regression model predictions
- Cost history and trends
- LLM-assisted predictions
Max
Potestas
High volume with full API access.
- Everything in Pro
- At least 200x more analyses than free - even during peak demand
- Full API access
- CI gating in your pipeline
- Priority model updates
Team
Amplitudo
For teams tracking AI spend together.
- Everything in Max
- Pooled team usage - no individual caps
- Team dashboards
- Admin controls
- Minimum 2 seats, up to 25
Max and above include the GitHub Action CI gate. See how CI gating works
Enterprise
SummaCustom contracts for large deployments. Custom contracts, PAYG API pricing, SSO, SLAs, and dedicated support for teams that need it all.
- Unlimited analyses
- PAYG API pricing
- SSO and SLAs
- Dedicated support
- Custom contract
Compare features across every plan
| Feature | Free Initium | Pro Certitudo | Max Potestas | Team Amplitudo | Enterprise Summa |
|---|---|---|---|---|---|
| Cost forecasting | |||||
| Token-exact input counts tiktoken offline for OpenAI, provider APIs for Claude and Gemini | |||||
| Heuristic + regression output prediction Local prediction using the fitted regression forest | |||||
| LLM-assisted prediction (Precise mode) GPT-5.4 Nano forecasts your output token count | − | ||||
| Bayesian P10 to P90 confidence band Honest uncertainty range under the headline cost | − | ||||
| Estimation surfaces | |||||
| Web estimator (paste a prompt) | |||||
| Multi-model calculator | |||||
| Workload calculators (chatbot, RAG, etc.) | |||||
| Multi-turn session simulator Conversation growth, context compounding, agentic tools | − | ||||
| Context file analyser Upload a long context file, see per-call burn | − | ||||
| Prompt compression Mini model rewrites your prompt; you only pay if shorter | − | − | |||
| API and integrations | |||||
| Public REST API POST /api/v1/estimate with a Bearer key | − | ||||
| API key issuance | − | Up to 10 | Up to 10 | Pooled | Custom |
| CLI (npx calcis) | |||||
| Browser extension overlay | − | ||||
| VS Code extension | |||||
| GitHub Action Free for public repos always | Public repos | ||||
| Embeddable cost badges | |||||
| Quotas | |||||
| Estimates per week Resets Monday 00:00 UTC. Soft cap on Free. | Soft cap | 65 | 650 | 6,500 pooled | Unlimited |
| Quota top-ups via credit packs | |||||
| Team and support | |||||
| Pooled team quota | − | − | − | ||
| Team admin dashboard | − | − | − | ||
| Seats included | 1 | 1 | 1 | 2 to 25 | Custom |
| SSO and SLAs | − | − | − | − | |
| Support | Community | Priority email | Priority email | Dedicated | |
Pricing FAQ
Quick answers on quotas, cancellation, refunds, and discounts. Anything else, the support address is at the bottom of every page.
What counts as an estimate?
One call to the estimator (web, CLI, browser extension, or REST API) that returns a token count and cost. Token counters and read-only calculator views do not count. The same prompt re-estimated against a different model counts as one extra estimate, because each model needs its own tokeniser pass.Can I cancel anytime?
Yes. From your dashboard, click Manage subscription. The Stripe customer portal handles cancellation, plan switches, and invoices. Your tier stays active until the end of the current billing period; no prorated refund for unused days, no charge after that period closes.What happens to my API key if I downgrade?
The key is preserved but disabled server-side. Any request authenticated with it returns 403 with a message naming the downgrade. Re-upgrading reactivates every key you had previously, no re-issuance needed. If you cancel entirely (back to Free), you can still see the keys in your dashboard but cannot use them until you upgrade again.Do you offer student or open-source discounts?
Open-source: yes. If your project is public on GitHub or another forge with an OSI-approved licence, email calcis.dev@gmail.com with the repo link and we will issue a 50% discount on Pro or Max for as long as the project stays open. Students: not yet a formal programme, but the same email gets you the same answer if you describe the use case.Is there a free trial?
Not a separate trial. The Free tier is the trial: every prediction layer except the LLM-assisted predictor (Pro and above) is available without an account. If you want to test the LLM predictor before committing, upgrade to Pro and cancel within 30 days for a full refund (see below).What is the refund policy?
First Pro or Max invoice is fully refundable within 30 days, no questions asked. Subsequent invoices are prorated to the unused portion of the billing period. Team-tier refunds work the same way per seat. To request, email calcis.dev@gmail.com with the Stripe receipt; refunds land back on the original card within five business days.
Token counts you can trust, updated every time a provider changes their pricing.
Cancel anytime from your dashboard. No contracts, no questions.
Billing handled by Stripe. We never see or store your card details.
Running low? Buy top-up analyses.
One-time packs that expire 90 days after purchase. Credits kick in automatically when you hit your plan quota.
About $0.040 per analysis.
- 50 analyses
- Kicks in when your plan is full
- Expires 90 days after purchase
About $0.030 per analysis.
- 200 analyses
- Kicks in when your plan is full
- Expires 90 days after purchase
About $0.024 per analysis.
- 500 analyses
- Kicks in when your plan is full
- Expires 90 days after purchase