Pay for the compute we route. No markup.
Every cent is the upstream provider cost. The plan price covers the router, observability, and budget guardrails — not a margin on tokens.
Free
- 1 API key
- OpenAI-compatible endpoint
- All five tiers (T0 → T4)
- Per-query telemetry
- Community support
Pro
- 10 API keys
- Per-domain calibration
- Usage exports (CSV / Parquet)
- Email support, 24h response
- Custom tier mapping
Scale
- Unlimited keys
- Bring-your-own provider keys
- Private deployment / VPC
- SLA + dedicated support
- Architecture review
How it compares
Entry pricing across major routing & observability tools.
Common questions
How do you compute "routed compute"?+
The actual upstream cost charged by OpenAI / Anthropic / Google for each request, summed across the month. We log it per query and you can see it on the Usage page.
Can I bring my own provider keys?+
On the Scale plan, yes — the proxy uses your keys and we only meter calls. On Free and Pro, requests go through Iudex Route's pooled keys.
What happens when I hit my quota?+
New requests return 402 until the start of the next UTC month. The router also starts downgrading tier picks as you approach the cap, so you rarely hit a hard wall.
Is there a free trial of Pro?+
The Free plan covers most prototype workloads. Upgrade only when you need the extra quota, exports, or per-domain calibration.
How does Iudex Route differ from OpenRouter or Helicone?+
OpenRouter is a billing gateway — you still pick the model. Helicone is observability — you still pick the model. Iudex Route picks the model for you based on a difficulty score, then logs the result.