---

name: ingestion-documents description: "Use R2R document ingestion, metadata, filters, exports, and collection/document lifecycle workflows." disable-model-invocation: true metadata: disco-role: operating license: MIT

Ingestion Documents

Use this sub-skill when the user needs to ingest files, raw text, chunks, or S3 content; edit metadata; build filters; or manage documents, chunks, and collections.

What it owns

• documents.create(...) with file_path, raw_text, chunks, or s3_url
• ingestion mode and orchestration choices
• document list/retrieve/download/export/delete flows
• metadata replace/append and delete-by-filter flows
• collection membership, document lists, and collection extraction setup
• chunk and collection listing that is part of ingestion lifecycle work

Start here

from r2r import R2RClient

client = R2RClient(base_url="http://localhost:7272")
result = client.documents.create(raw_text="Hello from R2R", metadata={"title": "demo"})
print(result.results)

Route out when the work becomes another topic

• Search, retrieval, or RAG on ingested content: ../retrieval-rag/SKILL.md
• Graph extraction or graph CRUD after ingestion: ../graph-workflows/SKILL.md
• Server/provider setup needed to make ingestion work: ../server-configuration/SKILL.md

Bundled assets

• references/document-workflows.md
• references/data-formats-and-filters.md
• references/troubleshooting.md
• scripts/ingestion_payload_builder.py