Sign in
manual-ingest

Ingest Manual Document

Ingest a single document manually.

The document will be processed through the full pipeline:

  1. Stored in PostgreSQL
  2. Chunked into smaller pieces
  3. Embedded with OpenAI embeddings
  4. Entities and relations extracted
  5. Knowledge graph updated in Neo4j

This is useful for:

  • Testing with real transcripts
  • Importing external documents
  • Quick demos without OAuth setup
post/v1/ingest/manual/document

Request body

workspace_idstring uuid required

Target workspace

vault_idstring uuid nullable

Optional vault to associate with

contentstring required

The text content to ingest

titlestring required

Document title

document_type'transcript' | 'notes' | 'document' | 'article' | 'email' | 'chat'

Types of manually ingested documents.

author_namestring nullable

Author name

author_emailstring nullable

Author email

source_urlstring nullable

Original source URL if applicable

created_atstring date-time nullable

Original creation date

tagsstring[]

Tags for categorization

metadataobject

Additional metadata

Response

Successful Response

document_idstring uuid required
statusstring required
messagestring required
processing_job_idstring uuid nullable

Changes

No recorded changes to this endpoint across all 1 revision of this API.