RAG
POST /rag/upload
Upload and index documents for retrieval-augmented generation
POST
Overview
Uploads one or more files, extracts their text, splits it into chunks, generates embeddings, and stores everything in the workspace’s RAG index so it can be queried via/rag/query.
Authentication
string
required
Bearer token from an authenticated Supabase session. See the note on RAG auth.
Request
Content type:multipart/form-data
File[]
required
One or more files to upload, as
files[] form fields.string
required
UUID of the workspace to associate the documents with. The workspace’s RAG index is auto-provisioned on first upload if it doesn’t already exist.
Supported File Types
Documents are split into ~1000-character chunks along sentence boundaries. CSV files are converted to structured text (with column headers preserved) before chunking.
Response
string
"success"string
e.g.
"Processed 3 files"object
string
string
ISO 8601
Errors
Partial success is allowed — some files can fail while others succeed. Check
results.details[].status for each file.