Start
Launch key
No hard-coded public price
The first step: get a Spectorn key, verify /v1/scan, then call a hosted Spectorn model.
- Gateway auth and tenant isolation
- Fail-closed scan before model calls
- Hosted models use Spectorn balance
Pay for the way you actually route models.
Spectorn sells one protected model surface. Hosted credit is the default path; private provider keys and dedicated contours are available for advanced deployments.
Private routes
For teams bringing their own endpoint
Spectorn protects and routes the request while hosted model cost is billed through your Spectorn credits, or through your own direct endpoint when configured.
- Saved provider for private contours
- Provider keys encrypted at rest
- Good for fast pilots
Hosted prepaid
From signed price config
When hosted providers are enabled, the gateway reserves balance on entry and settles the delta after actual usage is known.
- Reserve-then-settle accounting
- Low-cost model allow-list
- Overspend protection
Enterprise
Contact us
For dedicated VPS/clusters, on-premise, special provider routes, and data residency requirements.
- Self-hosted or dedicated contour
- Protection policies for your scenarios
- Pilot and responsibility boundaries before launch
How metering works
The homepage should not go stale every time model markets move. This page explains the mechanics; exact numbers come from Spectorn's signed price config or your private-route configuration.
Private provider
- What is charged
- Private endpoint cost, when configured
- Notes
- Spectorn does not pretend to resell someone else's tokens; gateway fee is configured separately.
Spectorn gateway
- What is charged
- Policy, scanning, ledger, memory
- Notes
- One API key, tenant isolation, scan before forward, and output scan.
Hosted provider credit
- What is charged
- Prepaid ledger when enabled
- Notes
- Reserve on entry, settle after response, abort when balance is exhausted.
Enterprise contour
- What is charged
- Contract terms
- Notes
- Hosting, load, providers, and data requirements are documented separately.
| Route | What is charged | Notes |
|---|---|---|
| Private provider | Private endpoint cost, when configured | Spectorn does not pretend to resell someone else's tokens; gateway fee is configured separately. |
| Spectorn gateway | Policy, scanning, ledger, memory | One API key, tenant isolation, scan before forward, and output scan. |
| Hosted provider credit | Prepaid ledger when enabled | Reserve on entry, settle after response, abort when balance is exhausted. |
| Enterprise contour | Contract terms | Hosting, load, providers, and data requirements are documented separately. |
Final public prices are published only after the commercial config is approved. Until then, this page avoids promising a specific RUB/USD tariff or showing stale model prices.
Start with a key, not a full infrastructure migration.
Verify the scan path, choose a Spectorn model, and move OpenAI-compatible traffic through Spectorn.