cURL
curl --request POST \ --url https://api.veryfront.com/projects/{project_reference}/environments/{environmentName}/search \ --header 'Authorization: Bearer <token>' \ --header 'Content-Type: application/json' \ --data ' { "vector": [ 123 ], "dimension": 768, "limit": 10, "threshold": 0.7 } '
{ "data": [ { "chunk": { "id": "<string>", "file_path": "<string>", "file_id": "<string>", "file_version_id": "<string>", "index": 123, "content": "<string>", "start_offset": 123, "end_offset": 123, "token_count": 123, "metadata": {} }, "score": 123, "match_type": "semantic" } ], "page_info": { "self": "<string>", "first": null, "next": "<string>", "prev": "<string>" }, "search_time_ms": 123, "total_matches": 123 }
Search using a pre-computed embedding vector.
Authentication via JWT token or API key. JWT: Get from Veryfront dashboard. API Key: Format "vf__" - create via /api-keys endpoint.
Query embedding vector
768 - 4096
Vector dimension (768, 1024, 1536, 3072, 4096)
768
1 <= x <= 100
0 <= x <= 1
Search results
Show child attributes