Pricing
Pay per seat. Not per call.
Helmstack charges for the humans on the deck — not the agents you connect, the tools you wire up, or the runs they make.
Solo
$0forever
Run a single agent for your own work. Everything except the bigger fleet limits.
Start free- 1 connected agent
- Up to 3 connectors
- 30-day activity log
- Email approvals
- Community Discord
Most teams pick this
Team
$49per seat / mo
When the side project becomes a real fleet. Bring your colleagues onto the deck.
Start 14-day trial- Up to 25 agents
- Unlimited connectors
- 90-day activity log
- Slack & email approvals
- Role-based access
- Audit log export
- Priority support
Enterprise
Customlet's talk
Self-hosted or VPC, SSO, audit, custom retention, dedicated support.
Talk to sales- Unlimited agents
- SAML SSO + SCIM
- Self-hosted or VPC
- Custom data retention
- Bring-your-own LLM provider
- Dedicated security review
- 99.95% SLA
Frequently asked
- What counts as an "agent"?
- A distinct workspace or runtime registered with Helmstack — a Claude Code instance, an OpenAI Assistant, a homegrown HTTP worker. Multi-skill agents count once.
- Do you charge per LLM call?
- No. You bring your own LLM provider keys. Helmstack never sits in the inference path unless you ask it to.
- Is there a free tier for hobby projects?
- Yes — Solo is free forever for one agent, no credit card required.
- Can I self-host?
- On Enterprise, yes. You get a Helm chart, a Terraform module, and a real engineer on speed dial.