outbound-agents

Propose better prompts from the agent's feedback, streaming the work

The same rewrite as proposing better prompts, as a stream of Server-Sent Events: a step frame for each piece of work once it is done (reading what people said needs work, then each prompt it rewrote), each with its measured duration, then one result frame whose data is the proposal, the before and after for each prompt. Nothing is written until the proposal is accepted. Spends one of the workspace's daily assistant turns.

Failures arrive as an error frame, because the stream has already answered 200: reason not_found when the agent is not this workspace's, limit when the workspace has used today's AI Assistant allowance, paused when the AI Assistant is paused, and no reason for anything else, including no feedback to learn from, whose sentence says so.

Closing the connection does not stop the rewrite.

post/outbound-agents/{agent_id}/proposals/stream/

Request

  • Base URL: https://api.getoneprofile.ai (the document declares no server; this is the origin it was published from)
  • URL: https://api.getoneprofile.ai/outbound-agents/{agent_id}/proposals/stream/
  • Auth: HTTP bearer

Path parameters

agent_idstring uuid required

Request body

also_slotsstring[] nullable

Teach the same feedback to these prompts too, even though none was given on their messages. Feedback names the one message it was given on, so without this a note on the opener can only ever change the opener. Each prompt is still rewritten against its own wording.

revision_requeststring

What to change about a proposal already seen.

slot_keysstring[] nullable

Only rewrite these prompts; every prompt that has feedback when null.

Response

A stream of Server-Sent Events; the final result frame's data is the proposal.

Changes

No recorded changes to this operation across all 1 revision of this API.