Agentset icon

Agentset

Interact with the Agentset API

Actions18

Ingest Job → Create

AI-generated

Summary

Create a new ingest job within a specified namespace to process and ingest various types of content payloads such as text, file URLs, web crawls, or YouTube videos.

Inputs

  • Namespace ID (required) — The unique ID of the namespace where the ingest job will be created (prefixed with ns_).
  • Payload Type (required) — Specifies the type of content payload to ingest; options are Text, File URL, Crawl, or YouTube.
  • Text (required) — The text content to ingest (required if Payload Type is Text).
  • File URL (required) — The URL of the file to ingest (required if Payload Type is File URL).
  • URL to Crawl (required) — Starting URL to perform a web crawl from (required if Payload Type is Crawl).
  • YouTube URLs (required) — Comma-separated list of YouTube video, channel, or playlist URLs to ingest (required if Payload Type is YouTube).
  • Additional Fields — Optional configuration including chunk size, external ID, file name, language code, ingest job name, processing mode, and optional tenant ID header.
  • Crawl Options — Options specific to Crawl payloads such as max depth, limit of results, and path filters.
  • YouTube Options — Options specific to YouTube payloads such as transcript languages and whether to include video metadata.

Output shape

a single record containing the created ingest job's details

The output is the JSON response from the Agentset API representing the created ingest job. Error cases return error messages if 'continueOnFail' is enabled.

Links

Discussion