The endpoints documented here let you ingest knowledge, search it with AI, and generate new content—all from your own code.
What you can do
• Upload raw text or files (PDF, DOCX, Markdown)• Organise knowledge with categories and topics
• Ask natural-language questions and get cited answers (
/search
)• Generate new text such as summaries or replies (
/generate
)
Everything is exposed through a RESTful JSON API.
Base URL
Authentication (API key)
Every request must include your organisation’s API key:Content type
Most requests/responses use JSON:multipart/form-data
is used only when uploading or updating a file (/content/file
).
Pagination
List endpoints share the same optional query parameters:Parameter | Default | Max | Notes |
---|---|---|---|
limit | 10 | 100 | items returned per page |
offset | 0 | — | items to skip (0-based) |
GET /categories?limit=20&offset=40
Error model
All errors follow one shape:• 201 Created – new resource written
• 202 Accepted – async processing started
• 204 No Content – success without body
• 400 Bad Request – malformed input
• 401 Unauthorized – missing / bad key
• 403 Forbidden – key lacks permission
• 404 Not Found – resource id unknown
• 409 Conflict – duplicate name, etc.
• 500 Internal Server Error
Using the reference
Each endpoint page includes:- Method & path
- High-level summary
- Detailed description
- Parameters & schema links
- Copy-paste examples for cURL & JSON
/categories
, /topics
), then upload some content (/content
) and try a quick question with /search
.
Happy building!