CAcasemark
Vaults

Ingest vault object

Changed on

Triggers ingestion workflow for a vault object to extract text, generate chunks, and create embeddings. For supported file types (PDF, DOCX, PPTX, XLSX, TXT, RTF, XML, HTML, Markdown, CSV/TSV, JSON/YAML/TOML, common source code files, ZIP, audio, video), processing happens asynchronously. ZIP archives always return a processing response, are unpacked recursively up to 5 levels, and each extracted file is created as an independent vault object and ingested via the normal pipeline. For unsupported types (images, etc.), the file is marked as completed immediately without text extraction.

post/vault/{id}/ingest/{objectId}

Request

  • Base URL
    https://api.case.dev
  • URL
    https://api.case.dev/vault/{id}/ingest/{objectId}
  • Auth
    HTTP bearer

Path parameters

idstring required

Vault ID

objectIdstring required

Vault object ID

Request body

callback_urlstring uri

Optional callback URL for asynchronous workflow completion.

page_boundariesinteger[]

Optional PDF pages that must begin a new chunk segment. Overlap never crosses these boundaries.

Response

Ingestion workflow started successfully

objectIdstring required

ID of the vault object being processed

workflowIdstring nullable required

Workflow run ID for tracking progress. Null for file types that skip processing.

status'processing' | 'stored' required

Current ingestion status. 'stored' for file types without text extraction (no chunks/vectors created).

queuedboolean

True when the document is waiting for ingestion capacity. It starts automatically; workflowId is null until then.

messagestring required

Human-readable status message

Changes