[/PRICING]

USAGE-BASED.
NO SURPRISES.

Pay for traces processed, not seats. Cap spend per agent. Self-host if your CISO insists.

DEVELOPER
$0/ forever

For prototyping and small projects.

  • [+]10k traces / month
  • [+]7-day retention
  • [+]1 project, 3 seats
  • [+]Community support
START_FREE
TEAM// MOST_POPULAR
$499/ month

For teams shipping agents to production.

  • [+]1M traces / month included
  • [+]30-day retention
  • [+]Unlimited projects, 25 seats
  • [+]Eval suite & policy engine
  • [+]SSO (SAML, OIDC)
  • [+]Email support, 1-day SLA
START_TRIAL
ENTERPRISE
Custom

For regulated industries and global scale.

  • [+]Unlimited traces
  • [+]Self-hosted (VPC or on-prem)
  • [+]BYO KMS keys
  • [+]SOC 2 Type II report
  • [+]Dedicated solutions engineer
  • [+]Custom SLAs, 24/7 on-call
CONTACT_SALES
[FAQ]

QUESTIONS WE GET A LOT.

What counts as a trace?

One full agent execution — from initial prompt through every reasoning step, tool call, and final response. Sub-spans are free.

Do you store our prompt data?

Only what you send. PII redaction runs at ingest and you can opt out of payload storage entirely (metadata-only mode).

Can I cap spend per agent?

Yes. Set hard ceilings per agent, per team, or per environment. Hits trigger configurable behavior: block, alert, or fall back to a cheaper model.

Self-hosted?

Available on Team (with annual commit) and Enterprise. Deploys via Helm into your VPC. We never see your traces.

TALK_TO_US →