> ## Documentation Index
> Fetch the complete documentation index at: https://velt-mintlify-e6426361.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Ingest memory knowledge (v2)

> Submit a new knowledge source for ingestion into Velt Memory using the v2 REST API to extract, chunk, and index content for grounded AI answers and judgments.

Use this API to submit reference material (guidelines, standards, policy docs) to your Memory knowledge base. Ingestion is asynchronous: the call returns immediately with a `sourceId`, and conversion, rule extraction, and embedding run in the background. Poll [Get Ingest Status](/api-reference/rest-apis/v2/memory/knowledge/ingest-status) until the status is terminal. Supported file types: PDF, CSV, Excel (`.xlsx`), and plain text.

# Endpoint

`POST https://api.velt.dev/v2/memory/knowledge/ingest`

# Headers

<ParamField header="x-velt-api-key" type="string" required>
  Your API key.
</ParamField>

<ParamField header="x-velt-auth-token" type="string" required>
  Your [Auth Token](/security/auth-tokens).
</ParamField>

# Body

#### Params

The request is a discriminated union on `source`: send `source: "inline"` with a base64 `file` (decoded size up to 5 MB), or `source: "fileRef"` with a `gs://` URI from [Get Upload URL](/api-reference/rest-apis/v2/memory/knowledge/upload-url) (up to 30 MB).

<ParamField body="data" type="object" required>
  <Expandable title="properties">
    <ParamField body="source" type="string" required>
      `inline` or `fileRef`.
    </ParamField>

    <ParamField body="file" type="object">
      Required when `source` is `inline`. `{ base64, mimeType, fileName, fileSize }`. `base64` is the encoded file bytes (decoded size up to 5 MB); `fileName` is 1 to 255 characters; `fileSize` is in bytes.
    </ParamField>

    <ParamField body="fileRef" type="string">
      Required when `source` is `fileRef`. The `gs://` URI returned by Get Upload URL.
    </ParamField>

    <ParamField body="mimeType" type="string">
      Required when `source` is `fileRef`. One of the supported types: `application/pdf`, `text/csv`, `application/vnd.openxmlformats-officedocument.spreadsheetml.sheet`, or `text/plain`.
    </ParamField>

    <ParamField body="organizationId" type="string">
      Scope the source to one organization.
    </ParamField>

    <ParamField body="documentId" type="string">
      Scope the source to one document. Requires `organizationId`.
    </ParamField>
  </Expandable>
</ParamField>

## **Example Requests**

#### Inline (file up to 5 MB)

```JSON theme={null}
{
  "data": {
    "source": "inline",
    "file": {
      "base64": "JVBERi0xLjQK...",
      "mimeType": "application/pdf",
      "fileName": "brand-guidelines.pdf",
      "fileSize": 184320
    }
  }
}
```

#### By reference (file up to 30 MB)

```JSON theme={null}
{
  "data": {
    "source": "fileRef",
    "fileRef": "gs://bucket/path/original.pdf",
    "mimeType": "application/pdf"
  }
}
```

# Response

Ingestion runs in the background. Poll Get Ingest Status with the returned `sourceId` until the status is `completed` or `failed`.

#### Success Response

```JSON theme={null}
{
  "result": {
    "status": "processing",
    "sourceId": "src_9a8...",
    "message": "Ingestion started."
  }
}
```

#### Failure Response

```JSON theme={null}
{
  "error": {
    "message": "ERROR_MESSAGE",
    "status": "INVALID_ARGUMENT"
  }
}
```

<ResponseExample>
  ```js theme={null}
  {
    "result": {
      "status": "processing",
      "sourceId": "src_9a8...",
      "message": "Ingestion started."
    }
  }
  ```
</ResponseExample>