# VectorSearchHit

Model ID: `VectorSearchHit`

One ranked chunk, carrying enough metadata to render without a second fetch.

Type: `object`

## Fields

| Field | Type | Required | Format | Allowed values | Default | Nullable | Description |
|---|---|---:|---|---|---|---:|---|
| `chunk_index` | `integer|null` | No |  |  |  | Yes | 0-based position of this chunk within its document. |
| `collection_group` | `string|null` | No |  |  |  | Yes | Corpus label the chunk belongs to. |
| `doc_id` | `string|null` | No |  |  |  | Yes | Identifier of the document this chunk came from. |
| `id` | `string` | Yes |  |  |  | No | Chunk primary key, '<doc_id>#<chunk_index>'. |
| `language` | `string|null` | No |  |  |  | Yes | Document language. |
| `metadata` | `object` | No |  |  |  | No | Per-source extras (topic ids, MeSH headings, payer, policy number, ...). |
| `network_id` | `string|null` | No |  |  |  | Yes | Stedi payer id, for policy chunks; empty when it could not be resolved. |
| `payer_key` | `string|null` | No |  |  |  | Yes | Payer registry slug the chunk belongs to; empty for consumer-health chunks. |
| `score` | `number|null` | No |  |  |  | Yes | Cosine similarity to the query; higher is closer. |
| `source` | `string|null` | No |  |  |  | Yes | Pipeline that produced the chunk. |
| `tags` | `array` | No |  |  |  | No | Health tags derived from the document's own metadata. |
| `text` | `string|null` | No |  |  |  | Yes | The chunk body that was embedded. |
| `title` | `string|null` | No |  |  |  | Yes | Document title. |
| `url` | `string|null` | No |  |  |  | Yes | Source URL of the document, when known. |

## Example

None declared in canonical OpenAPI.