> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sundaypyjamas.com/llms.txt
> Use this file to discover all available pages before exploring further.

# RAG API

> Upload documents and query them with retrieval-augmented generation

## Overview

The RAG API lets a workspace ingest documents, chunk and embed them, and query them with natural-language questions that return a generated answer plus the source passages it was grounded in.

<Steps>
  <Step title="Upload documents">
    POST files to [`/api/v1/rag/upload`](/api-reference/rag/upload). Each file is text-extracted, chunked, embedded, and indexed per-workspace.
  </Step>

  <Step title="Query">
    POST a natural-language question to [`/api/v1/rag/query`](/api-reference/rag/query). The engine retrieves the most relevant chunks and returns a generated answer with cited sources.
  </Step>
</Steps>

## Base URL

```
https://suite.sundaypyjamas.com/api/v1/rag
```

<Warning>
  Unlike the rest of the platform API, the RAG endpoints currently authenticate via a **Supabase session** (the same auth used by the AI Suite web app) rather than a workspace API key. If you're integrating from a backend service rather than the AI Suite frontend, reach out about API-key support for these endpoints before building against them in production.
</Warning>

## Quotas

Each workspace has a monthly query quota (`rag_workspaces.monthly_quota`, defaults to 1,000). Queries increment usage by 1; uploads are not quota-limited but are capped at 100 MB per file.

## Resources

<CardGroup cols={2}>
  <Card title="Upload" icon="upload" href="/api-reference/rag/upload">
    Ingest documents (PDF, DOCX, TXT, MD, CSV, JSON) into the workspace's RAG index.
  </Card>

  <Card title="Query" icon="magnifying-glass" href="/api-reference/rag/query">
    Ask questions and get grounded answers with source citations.
  </Card>
</CardGroup>