Usage & Tokens Policy
Effective date: July 21, 2026
Sevoria meters usage in tokens, the actual units of work the AI does. Every reply shows its real token cost underneath, so what you see is exactly what was spent. A short answer is a few hundred tokens; a long code generation can be several thousand.
How tokens work
- Cost is based on the work actually done: mostly the length of the answer, plus a small share of the reading. Bigger models (Pro, Max) weigh more per token because they cost more to run.
- Failed or stopped responses are not charged. If it errors out, or you stop it before it answers, that turn is free.
- Regenerating an answer is new work and costs tokens again.
Your allowances
Every plan gets a weekly token allowance, and paid plans also get a 5-hour window of premium model time. Both reset on their own, always.
| Plan | Weekly tokens | Premium time (per 5 hours) |
|---|---|---|
| Free | 250,000 | Core & Reason & Code only |
| Pro | 1,000,000 | 75,000 tokens of Pro |
| Max | 3,000,000 | 250,000 tokens of Pro & Max |
- When your 5-hour premium window is used up, Sevoria keeps answering with Core instead of stopping, and tells you it switched. The window reopens on its own.
- The weekly allowance is the only hard limit. If you reach it, Sevoria tells you exactly when it resets.
Where to see usage
- Under each reply: the token cost of that answer.
- In chat: the usage meter shows both windows and their reset times.
- Settings → Billing: weekly usage, recent activity, and billing history.
Resets, expiry, transfers
- Windows reset automatically; unused allowance does not roll over.
- Tokens have no cash value, can't be transferred between accounts, and are forfeited if an account is deleted or terminated for abuse.
Billing questions: Refund & Credits Policy. Plan details: Pricing.
Questions? billing@sevoria.co · Operator: Sevoria · Jurisdiction: the operator's principal place of residence.