PRICING

Signal Before Action.

Infrastructure-grade scoring for automated decision workflows.

Deterministic, calibrated, versioned scoring primitives accessible via a simple API.

Back to Portal Open Swagger

Starter

$0

  • 1,000 scoring calls per month
  • 5 requests per second
  • healthcare.denial workflow
  • Deterministic scoring + explanations
  • Versioned model responses

Best for testing and integration.

Growth

$0.003 / scoring call

  • First 1,000 calls/month included ($0)
  • $0.003 per call after included usage
  • Production rate limits
  • Monthly billing via Stripe metering
  • Self-serve checkout in portal dashboard
  • Access to public workflows

Best for early production workloads and predictable pay-as-you-go usage.

Pro

$99 / month + usage tiers

  • Includes first 50,000 scoring calls/month
  • $0.0020 per call from 50,001 to 250,000
  • $0.0015 per call above 250,000
  • Monthly billing via Stripe metering
  • Lower effective unit cost at sustained volume

Designed for steady production traffic and lower blended per-call cost.

FAQ

How do I start for free?
Mint an instant no-signup trial key with POST /v1/keys/trial, then call POST /v1/score and inspect usage with GET /v1/limits / GET /v1/usage.
How do I start paid billing?
Use /portal/dashboard to start Stripe Checkout (growth/pro), then create or rotate API keys in the same workspace session. Trial keys can be upgraded in place via POST /v1/billing/checkout-session (Bearer trial key). Admin automation can also use POST /v1/billing/checkout/session and POST /v1/apikeys/issue.
How are Growth and Pro billed?
Both use Stripe metered billing. Growth is pay-as-you-go after the first 1,000 monthly calls. Pro adds a $99 monthly commit with lower overage rates at higher volume.
What is a scoring call?
One successful 2xx response from a scoring endpoint. POST /v1/score is preferred, and legacy scoring endpoints count toward the same meter.
How do I check usage?
Use GET /v1/limits for envelope status and GET /v1/usage for billable usage totals.
How do I report real-world mismatches?
Use POST /v1/feedback with the scoring response x-request-id; feedback telemetry calls are non-billable.
Are scores deterministic?
Yes. Identical inputs return identical outputs for a given model version.
Are models versioned?
Yes. Every response includes model_version.
Do you store submitted payloads?
No. Payloads are processed in-memory and not retained.
Can I pin a model version?
Yes. Version pinning is supported.