Start operating your agents with essential visibility, controls, and skills.
Start free- Up to 5 agents · 2 apps · 2 team members
- 10,000 runs and 200 credits every month
- Run history, tokens, and agent health
- Skills Hub and Testing Lab
Start free. Grow your agents, apps, and operating capacity.
Start with live operations and Testing Lab. Add feedback, quality reviews, and prompt approvals with Hobby, Team, or Scale.
| Features | Free | Hobby | Team | Scale |
|---|---|---|---|---|
| Credits & capacity | ||||
| Monthly credits | 200 | 2,000 | 6,000 | Custom |
| Welcome credits (first month only) | Not included | Not included | 1,000 | Custom |
| Pay-as-you-go | Not included | Up to $250/mo | Up to $1,000/mo | Custom |
| Connected agents | Up to 5 | Up to 15 | Up to 50 | More than 50 |
| Runs per month | 10,000 | 100,000 | 500,000 | Agreed plan |
| Connected apps | Up to 2 | Up to 5 | Up to 10 | More than 10 |
| Team members | Up to 2 | Up to 5 | Unlimited | Unlimited |
| Credits per lab test | By model, from 1 | By model, from 1 | By model, from 1 | Agreed plan |
| Default run retention | 30 days | 30 days | 90 days | Agreed plan |
| Provider token charges included | Not included | Not included | Not included | Discuss with us |
| Observe | ||||
| Run history, errors & payloads | Included | Included | Included | Included |
| Agent history logs | Included | Included | Included | Included |
| Error explanations & suggested fixes | Not included | Included | Included | Included |
| Token & cost report | Included | Included | Included | Included |
| Estimated costs & predictions | Not included | Included | Included | Included |
| Agent health & heartbeats | Included | Included | Included | Included |
| Agent, app & model usage breakdowns | Included | Included | Included | Included |
| Error grouping & failure analysis | Included | Included | Included | Included |
| Evaluate & improve | ||||
| Feedback scores | Not included | Included | Included | Included |
| Quality reviews | Not included | Included | Included | Included |
| Skills Hub | Included | Included | Included | Included |
| Upload skills | Not included | Included | Included | Included |
| Baseline / candidate lab tests | Included | Included | Included | Included |
| Before / after test results | Included | Included | Included | Included |
| Draft skills & review before applying | Included | Included | Included | Included |
| Operate | ||||
| Pause, resume & model overrides | Included | Included | Included | Included |
| Versioned prompts & approvals | Not included | Included | Included | Included |
| Stop & restart controls | Included | Included | Included | Included |
| Alert rules & thresholds | Not included | Included | Included | Included |
| Spend alerts & automatic pausing | Not included | Included | Included | Included |
| Email alert notifications | Not included | Included | Included | Included |
| Security & integration | ||||
| Signed commands & scoped keys | Included | Included | Included | Included |
| Command audit trail | Included | Included | Included | Included |
| Replay protection | Included | Included | Included | Included |
| SDK & HTTP integration | Included | Included | Included | Included |
| Webhooks & API callbacks | Included | Included | Included | Included |
| Prompt-generated integrations | Included | Included | Included | Included |
| Fail-open reporting | Included | Included | Included | Included |
| Node & Python SDKs | Included | Included | Included | Included |
| Two-factor sign-in (authenticator app) | Included | Included | Included | Included |
| Data export & deletion | Included | Included | Included | Included |
| Data Processing Addendum | Included | Included | Included | Included |
| ACP stores your provider keys | Not included | Not included | Not included | Discuss with us |
| Single sign-on (SSO / SAML) | Not included | Not included | Not included | Discuss with us |
| Support & guidance | ||||
| Onboarding meeting | Not included | Not included | Included | Included |
| Support response time | Within 1 day | Within 1 day | Within 1 day, answered first | Agreed with your team |
| Priority support | Not included | Not included | Included | Included |
| Weekly consultation meeting | Not included | Not included | Included | Included |
These are plan inclusions; new email delivery, skill upload, and pay-as-you-go services are subject to activation. Cost reports calculate usage from recorded tokens and provider rates; they are not provider invoices. Predictive forecasting is planned. Provider-key storage, single sign-on (SSO / SAML) and token coverage are optional Scale discussions, not standard features currently available.
Credits pay for the AI work ACP runs for you: Testing Lab tests, error explanations and suggested fixes, quality reviews, and test messages to an agent. Each action costs 1 credit per $0.005 of AI cost, at least 1. An error explanation or a quality review is usually 1 credit; a lab test is usually 1 to 9, depending on the model it replays (more on the largest models), and Testing Lab shows the estimate before you run it. Connecting apps, monitoring, the Skills Hub and the docs use no credits. Free includes 200 credits a month; unused credits do not carry over. Access to each feature depends on your plan.
The recommendation uses whichever count requires the higher plan. For example, 5 agents and 3 apps fit Hobby; 16 agents and 2 apps fit Team. Hobby is $50/month with 2,000 credits, Team is $150/month with 6,000 credits (plus 1,000 welcome credits in the first month), and Scale is agreed directly.
Model-provider usage is separate. ACP is a control plane; your agent still calls its model provider directly.
Request early access with your name, work email, and project. Already have an account? Use Sign in to open your workspace.