Retrieval and Documents

Orlo Platform supports task-scoped document ingestion and retrieval-backed inference.

Documents

Documents attach to tasks. Current supported ingestion modes include:

  • inline text
  • inline markdown
  • pre-chunked content
  • PDF ingestion with async extraction

Retrieval

Retrieval is task-scoped. Multiple documents under the same task can contribute context to one inference.

PDF flow

PDFs are handled asynchronously:

  1. upload or start direct upload
  2. finalize the document
  3. extract text
  4. chunk content
  5. embed chunks
  6. mark the document ready

Attribution

Retrieval-backed inference returns attribution in debug/audit contexts so you can inspect what source material contributed to the response.