Pricing
Self-host it free, let us run it for $99 a month, or deploy it in your own cloud. Every plan runs on the AI subscription you already pay for.
Cloud starts with a call · We set up your first goal with you, by hand
Open source
$0forever
Run the whole engine yourself, on your Mac or your own Linux server.
Cloud
$99/ month
Everything in open source, hosted and set up for you. Nothing to install.
Enterprise
Custom
Agentic engineering operations, deployed in your cloud with the controls your security team needs.
How Cloud onboarding works
Every Cloud team is onboarded by hand. You get a working goal in week one, not a blank workspace.
A short call about the outcome you want and the tools you use. We tell you honestly if it's a fit.
We connect your tools and AI plan, pick the right premade agent or write one, and agree the metric and target together.
Your agents work toward the goal. We review the results with you and tune the setup until the number moves.
Book a call7-day money-back guarantee once you start.
What you'll actually pay
No credits to count, no pool that runs dry mid-month, no marked-up tokens. Your Claude, ChatGPT or Gemini plan does the work, and your provider bills you directly.
Example monthly bill4 teammates · 5 goals
Compare plans
| Feature | Open source | Cloud | Enterprise |
|---|---|---|---|
| Teammates & goals | |||
| Crew teammates | Unlimited | Unlimited | Unlimited |
| Goals with measured targets | ✓ | ✓ | ✓ |
| Auto-improvement toward your goals | ✓ | ✓ | ✓ |
| Premade business agents | Rolling out | Rolling out | Rolling out |
| Engineering playbooks | ✓ | ✓ | ✓ |
| Custom playbooks built with you | — | — | ✓ |
| Channels & tools | |||
| Slack and WhatsApp | ✓ | ✓ | ✓ |
| Gmail updates and granted mailbox work | ✓ | ✓ | ✓ |
| Browser per teammate | Self-run | Hosted | Your cloud |
| MCP servers and API tools | ✓ | ✓ | ✓ |
| Custom integrations | — | — | ✓ |
| Control & security | |||
| Approvals and autonomy limits | ✓ | ✓ | ✓ |
| Encrypted secrets vault | ✓ | ✓ | ✓ |
| OS-enforced sandbox | ✓ | ✓ | ✓ |
| Run logs and cost per goal | ✓ | ✓ | ✓ |
| Roles and per-goal sharing | ✓ | ✓ | ✓ |
| SSO (SAML / OIDC) and SCIM | — | — | ✓ |
| Audit log export | — | — | ✓ |
| Deployment & support | |||
| Where it runs | Your Mac or server | AgentWorks cloud | Your VPC or private cloud |
| Onboarding | Docs | First goal set up with you | Scoped pilot |
| Support | Community | Priority email | Dedicated + SLA |
Your hosted workspace, unlimited Crew teammates and goals, a browser for every teammate, every premade agent, onboarding for your first goal, and priority support.
Yes. AgentWorks runs on the Claude, ChatGPT, Gemini or Cursor plan you already pay for, through that vendor's own coding agent (or its API key). We never mark up tokens, so your AI bill stays with your provider.
Light use works on a $20 plan. If teammates run all day, a $100–$200 plan (Claude Max or ChatGPT Pro) is the sweet spot. You can connect several and route each job to the one that fits.
The open-source edition is free forever. Cloud comes with a 7-day money-back guarantee: if it's not for you, cancel in the first week and we refund you in full.
Yes. Cancel from your workspace settings. You keep access until the end of the billing period and can export your goals, agents and logs.
Yes. It's the same engine, so goals, agents and playbooks move either way.
By deployment and scope, as an annual contract. Most teams start with a paid pilot on one engineering goal with agreed success metrics, then expand.
Pick a premade agent, connect your tools and set the target. Book a short call and we'll set up your first goal with you.
