PromptUI icon

PromptUI

List, fetch and execute PromptUI prompts via the public API

Prompt → Execute

AI-generated

Summary

Execute a selected PromptUI prompt by submitting its inputs and optional media, model overrides, and callback details. Return the queued job response immediately, or wait for completion and return the text or binary result.

Inputs

  • Prompt (required) — Select a prompt from the list or provide its ID or name.
  • Inputs (required) — Map values to the prompt’s non-media inputs; numeric inputs are numbers and other inputs are strings.
  • Binary Inputs — Map prompt image or video inputs to binary properties on the incoming item. Each mapped file is sent with its MIME type and base64 content.
  • Model Overrides — Optionally specify a model ID for one or more zero-based prompt step indexes. Overrides apply to this execution.
  • Wait for Completion — Defaults to false. When enabled, poll the job until it finishes, then return its result.
  • Poll Interval (Seconds) — When waiting, how often to check job status; defaults to 3 seconds.
  • Timeout (Seconds) — When waiting, maximum time before timing out; defaults to 900 seconds. A timed-out job may still finish and can be retrieved later with Job → Get Result.
  • Callback URL — Optionally provide an HTTPS URL for the server to notify when the job finishes. Requires callbacks to be enabled on the server.
  • Callback Payload — Optional payload echoed back in the callback for correlating the job; sent only when a callback URL is provided.

Output shape

One n8n item per incoming item: either the API job response, or the completed result when waiting is enabled.

Without waiting, the item’s JSON contains the API response, including the job ID and status. With waiting enabled, a text result is returned as JSON fields jobId, contentType, and result; a non-text result is attached as binary data under binary property data, with jobId and contentType in JSON. Failed jobs and wait timeouts raise errors.

Links

Discussion