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

# Create knowledge base

> Create a knowledge base and its vector collection

Creates a knowledge base your agents can retrieve answers from. Creating it also
provisions the underlying vector collection, so this is the first step before
adding sources or uploading documents.

### Body parameters

<ParamField body="kbKey" type="string" required>
  Short key, unique within your tenant — e.g. `support`
</ParamField>

<ParamField body="kbName" type="string" required>
  Display name — e.g. `Support Knowledge Base`
</ParamField>

<ParamField body="description" type="string">
  What this knowledge base contains
</ParamField>

<ParamField body="retrievalScoreThreshold" type="number">
  Cosine relevance floor for retrieval, between `0` and `1`. Overrides the
  platform default — raise it to return fewer, more relevant chunks
</ParamField>

<ParamField body="metadata" type="object">
  Arbitrary key/value data stored with the knowledge base
</ParamField>

<CodeGroup>
  ```bash curl theme={null}
  curl -X POST https://api.agentic-os.com/api/v1/knowledge-bases \
    -H "Authorization: Bearer $AGENTIC_OS_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{ "kbKey": "support", "kbName": "Support Knowledge Base" }'
  ```

  ```python Python theme={null}
  requests.post(
      "https://api.agentic-os.com/api/v1/knowledge-bases",
      headers={"Authorization": f"Bearer {API_KEY}"},
      json={"kbKey": "support", "kbName": "Support Knowledge Base"},
  )
  ```

  ```javascript Node theme={null}
  await fetch("https://api.agentic-os.com/api/v1/knowledge-bases", {
    method: "POST",
    headers: {
      Authorization: `Bearer ${API_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({ kbKey: "support", kbName: "Support Knowledge Base" }),
  });
  ```
</CodeGroup>

### Next steps

Add content with `POST /knowledge-bases/{kbId}/sources`, or upload a file to
`POST /knowledge-bases/{kbId}/documents/upload`, then queue ingestion with
`POST /knowledge-bases/{kbId}/ingestion-jobs`.