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

# Knowledge Base

> Upload documents and let Navigator retrieve relevant context during chat.

Navigator can use project documents as retrieval context. Uploaded documents are chunked, embedded, stored, and searched when the assistant needs grounded answers from your files.

## Retrieval flow

```mermaid theme={null}
flowchart LR
  Upload["Upload document"] --> Chunk["Chunk text"]
  Chunk --> Queue["Queue embedding jobs"]
  Queue --> Embed["Generate embeddings"]
  Embed --> Vector["Store vectors"]
  Question["User question"] --> Search["Vector search"]
  Vector --> Search
  Search --> Answer["Assistant answer with context"]
```

## Setup outline

<Steps>
  <Step title="Enable Supabase extensions">
    Run the project migrations that enable pgvector, pgmq, pg\_cron, and the RAG utility functions.
  </Step>

  <Step title="Deploy the embed function">
    Deploy the Supabase Edge Function used to generate embeddings asynchronously.
  </Step>

  <Step title="Configure secrets">
    Add Supabase credentials and any optional AI Gateway settings.
  </Step>

  <Step title="Upload and test">
    Upload a document, monitor embedding progress, then ask a question that should retrieve from that document.
  </Step>
</Steps>

## Search endpoints

```http theme={null}
POST /api/projects/{projectId}/search
GET /api/projects/{projectId}/status
POST /api/knowledge-base/search
```

<Tip>
  Use project-specific search when a thread should only draw from one archive or customer workspace. Use global search for broad personal knowledge lookup.
</Tip>