Build your first governed AI workflow
Start with governed LLM traffic, then add templates or teammates when your workspace is enabled.
1. Sign up
Create your account at dashboard.curate-me.ai/signup . Self-serve signup gives you gateway access for direct LLM calls.
2. Proxy your first LLM call
Point your existing SDK or HTTP client at the Curate-Me gateway and add your Curate-Me key. Start with one of these:
3. Choose the work pattern
After the gateway is working, choose how you want Curate-Me to run work:
| Pattern | Best for |
|---|---|
| Direct gateway traffic | Existing applications and agents that already call LLM providers |
| Autopilot templates | Repeatable work such as reports, research, code review, security audits, support, and finance |
| Teammates and managed runners | Private-beta agent teams that need tool access, files, browser automation, or desktop streaming |
See Autopilot Templates and Teammates and Roles for the full catalog.
4. Connect tools
Connect only the tools your workflow needs:
- Provider secrets for production gateway traffic
- Slack or Teams for notifications, commands, and approvals
- GitHub and email for development, support, and report workflows
See Slack and Teams and GitHub and email.
5. Review activity
Visit the dashboard to see:
- Gateway traffic — request IDs, providers, models, costs, denials, and retries
- Reports — Autopilot outputs and shareable HTML reports
- Cost tracking — real-time spend by org, key, model, tag, and runner session
- Audit log — every action, every decision, fully traceable
- Approvals — pending HITL requests waiting for your sign-off
Adjust budgets, PII scanning, rate limits, and model access at any time. See Adjusting Behavior.
For developers
If you want to integrate the gateway directly into application code or self-host the platform, start here: