Documentation is on its way
Full documentation for Orchestra — setup guides, agent configuration, workflow management, and API references — is currently being written.
In the meantime, the best way to get help is to join the Telegram community where the team and other users are active.
Quick start
1. Sign up at hiorchestra.com and connect your AI provider key (Anthropic, OpenAI, Google AI, or via OpenRouter).
2. Pick an agent role from the catalog and deploy it in one click.
3. Talk to your agent through the Orchestra portal, or connect a channel your team uses — Telegram, Slack, Microsoft Teams, and more.
4. Explore workflows and skills from the dashboard to automate recurring tasks.
Models and included inference
Orchestra is model-agnostic. Connect your own provider — Anthropic, OpenAI, Google, OpenRouter, xAI, DeepSeek or Moonshot — and set a default per agent. When you bring your own keys, tokens are billed to you directly by that provider at their price. Orchestra never marks them up and never takes a cut.
If you do not have a provider, inference is included in every plan so you can run agents from day one without opening an account anywhere else. Included inference is capped at USD 10 of provider cost per month under fair use, across all plans.
Reaching the cap does not switch anything off silently: you are notified before it is reached, and you can connect your own provider key at any time to remove the limit entirely. Bringing your own key is always the cheaper path at volume, which is why it stays the default.
Get help
Join the Orchestra Telegram community for support, tips, and updates: t.me/hiorchestra
For account or billing questions, email hey@hiorchestra.com