- 3 active agents
- 1,000 agent runs / month
- 10 prebuilt integrations
- Basic memory & run traces
- Visual workflows
- SSO & audit logs
- Unlimited agents
- 100,000 agent runs / month
- All 200+ integrations
- Visual workflows & guardrails
- Long-term memory + API access
- Priority support
- Everything in Pro
- Unlimited runs & seats
- SSO, SCIM & audit logs
- Self-hosted or VPC deployment
- SOC 2, custom security review
- Dedicated success manager
Questions, answered
Everything you need to know before you start. Still curious? Our team is one message away.
Is there really a free plan?
Yes. The Free plan is free forever — three active agents, 1,000 runs a month, and no credit card to sign up. It's a real way to evaluate Agentic in production-like conditions, not a 7-day trap.
What counts as an "agent run"?
A run is one full execution of an agent or workflow from trigger to completion — including all of its reasoning steps and tool calls. A scheduled job that fires hourly counts as one run per fire.
Can I bring my own model and API keys?
Absolutely. Connect your own provider keys, or use our managed routing. Agentic supports leading reasoning, fast, and vision models, and you can set a default per agent.
How do guardrails and approvals work?
Every agent can have budgets, step limits, and policy rules. Sensitive tools can require human approval before they fire, so an agent never takes an irreversible action on its own.
Is my data used to train models?
Never. Your prompts, memory, and tool data stay private to your workspace and are not used to train any model. Enterprise adds SSO, audit logs, and self-hosted deployment.
Can I switch plans or cancel any time?
Yes. Upgrade, downgrade, or cancel from billing settings whenever you like. Changes apply immediately and we prorate the difference automatically.
Start building agents free
Create your workspace in under a minute. Upgrade only when your agents are ready for production scale.