Skip to content

Artificial Intelligence · 1 min read

Deploy to API for Managed Agents

Deploy to API lets your own software talk to a published Managed Agent. Your website, app, or backend can start a conversation, send messages, and get the agent's replies, without building a separate agent or workflow.

How it works

  • Adding the Invoked via API trigger makes the agent available through the API.
  • The agent can keep its other triggers too.
  • API requests always use the agent's published version, so your app gets the version you've already tested.

Set it up

  1. Go to AI Agents > Agent Studio and open the Managed Agent.
  2. Do one of the following:
  3. In the Triggers row, add Invoked via API.
  4. Open the agent's ⋮ menu and choose Deploy to API.
  5. Review the endpoint and the sample requests in cURL, Python, or JavaScript.
  6. Create or choose an API key for your account. Store it securely on your server.
  7. Publish the agent.

Start and continue a conversation

  • Send a request to start a conversation. The response includes a conversation ID.
  • Send that conversation ID with later messages so the agent keeps the context.

Invoked via API or from a workflow?

  • Use Invoked via API when your own software starts the conversation.
  • Use Invoke Managed Agent in a workflow when something in your account, such as a form or tag, should start the agent.

Security tips

  • Never put your API key in browser code. Call the API from your server.
  • Rotate keys you think may be exposed.

Did this answer your question?

No, I need help

More in Artificial Intelligence