Pricing
Priced per seat, not per surprise
Every plan includes tracing, evals and guardrails. You are only ever paying for volume, retention and the people who configure it.
Starter
$0
forever
Everything you need to get one agent safely in front of real users.
50k traced spans / month
7-day trace retention
3 eval suites
Default guardrail pack
Community support
Pro
Most popular
$79
per seat / month
For teams whose customers now depend on the agent working.
2M traced spans / month
90-day retention
Unlimited evals and guardrails
CI checks and cost budgets
Slack and PagerDuty alerts
Priority support, 1 business day
Scale
Custom
billed annually
For regulated teams, very large volumes, or anything that needs to live in your own VPC.
Unlimited spans
Self-hosted or VPC deployment
SSO, SCIM and audit log
99.99% uptime SLA
Named support engineer
Custom data residency
Compare every plan
Feature
Starter
Pro
Scale
Traced spans / month
50k
2M
Unlimited
Trace retention
7 days
90 days
Up to 365 days
Eval suites
3
Unlimited
Unlimited
Guardrails
Default pack
Custom rules
Custom rules
CI integration
—
GitHub, GitLab
GitHub, GitLab
Cost and latency budgets
—
Included
Included
Alerting
Slack, PagerDuty
Slack, PagerDuty, webhook
Deployment
Cloud
Cloud
Cloud, VPC or self-hosted
SSO and SCIM
—
SSO
SSO and SCIM
Support
Community
1 business day
Named engineer
Billing questions
What counts as a traced span?
One span is a single unit of work inside a run — a retrieval call, a model call, a tool invocation. A typical agent run produces between four and twelve spans.
Can we change plans mid-cycle?
Do you charge for seats that only read?
What happens if we exceed our span limit?
Still deciding? Start on the free plan
No card, no call. Move up when your agent has real traffic and you need the retention.