Search project content
Vision API
Search
Hybrid semantic and visual search across a project drawing set.
POST
Search project content
Search answers “where does this appear?” by running hybrid semantic + keyword search across all indexed content in a project. Optionally combine with visual similarity search using a reference image.
Search is synchronous and returns results immediately (not an async job).
Text and visual results are merged and ranked by combined relevance score.
Request
| Field | Type | Required | Description |
|---|---|---|---|
query | string | Yes | Natural language search query |
limit | integer | No | Maximum results to return (default: 20) |
source_types | string[] | No | Filter by source type (block, feature, file) |
image_url | string | No | Storage URI (gs:// or s3://) for visual similarity search |
Visual Search
Whenimage_url is provided, search combines text-based results with visual similarity matching. The image must be a storage URI referencing a file within the target project.
Response
| Field | Description |
|---|---|
id | Composite identifier ({source_type}:{source_id}) |
content | Matched content text |
context | Human-readable location context |
score | Relevance score (higher is better) |
source_type | Type of content matched (block, feature, or file) |
source_id | ID of the matched entity |
Authorizations
API key prefixed with sk_. Example: Authorization: Bearer sk_xxx
Headers
API version
Path Parameters
Project ID
Body
application/json
Response
Search results.