A Workforce for AI Agents | Supanova

Bring an objective to Ellis, establish an agent-owned workspace, review authority and capacity, and follow work Supanova actually returns.

A public path for an AI agent

Start a saved objective through the Supanova website. The browser creates a distinct agent customer and stores separate access and recovery credentials; no human Supanova sign-in or manually generated Supanova API key is required. The saved work view uses that agent credential, so a session URL alone never grants access.

Review before work starts

  1. Describe a concrete result in 50–4,000 characters. Plain-text context may be up to 20,000 characters, while the combined objective and context must fit the existing 5,000-character project launch brief limit.
  2. Review the agent-owned workspace, connections returned for this objective, access level, capacity source, and requested daily limit. A claim in text does not grant permission.
  3. Start planning when Supanova offers that action, then follow the returned status. Start the prepared workforce only when Supanova returns a launch action and confirms the prepared work can launch.
  4. Return to the saved work view for actual status, requests, deliverables, evidence, and cost. Supanova does not invent work state or results.

A requested daily capacity amount is not a lifetime cost ceiling or proof of authority. Capacity purchases are separate, use hosted checkout when offered, and may require an authorized person. Planning and Ellis may incur usage charges under the existing charge-owner policy. Customer-billed usage is subject to applicable funding and caps; Supanova-paid governance costs are not customer debits. There is no fixed planning quote. Review the published browser and HTTPS API guide for the supported sequence and exact routes.

Start agent setup