Cloud API
Plans and quotas
Monthly call allowances per plan and how standard and neural calls are counted.
| Plan | Price | Standard calls / month | Neural calls / month | Data |
|---|---|---|---|---|
| Trial (7 days) | $0 | 5,000 | 500 | Zero data retention |
| Contributor | $0 | 10,000 | 500 | Anonymized telemetry helps improve zn |
| Starter | $29 / month | 500,000 | 5,000 | Zero data retention |
| Growth | $99 / month | 2,000,000 | 20,000 | Zero data retention |
| Enterprise | Custom | Custom | Dedicated | DPA, enterprise@usezn.com |
Need more in a given month? Credit packs: Flex $19 (100,000 calls), Pro $49 (500,000), Scale $99 (1,000,000). Current prices and annual billing: pricing.
How calls are counted
- Each
POST /analyzeor hosted MCPanalyzecall counts once, whichever layer decided. - Retries are new requests and count again.
- Local zn-gate checks (CLI, SDK, local MCP server) run on your machine and never count.
Your live usage is in the dashboard. The RateLimit-* headers on successful responses show what is left in the current window.