Semantic search | Dialogue API

Runs a vector similarity search over the corpus using a natural language query. Returns the most semantically relevant comments, ranked by cosine similarity. Optionally narrow results with metadata filters applied after the vector match.

Request

This endpoint expects an object.

querystringRequired

Natural language query to search for semantically similar comments.

limitintegerOptional1-50Defaults to 10

Number of results to return.

thresholddoubleOptional0-1Defaults to 0.4

Minimum cosine similarity score (0–1). Lower values return broader, less precise results.

emotional_registerenumOptional

The dominant emotional tone of a comment.

life_stageenumOptional

The inferred life stage of the commenter.

productstringOptional

Only return results mentioning this product.

Response

Semantically matched comments with similarity scores

resultslist of objects

Errors

400

Bad Request Error

500

Internal Server Error

$	curl -X POST https://api.dialogue-ai.co/api/v1/search \
>	-H "Content-Type: application/json" \
>	-d '{
>	"query": "worried about paying off mortgage before retirement",
>	"limit": 10,
>	"threshold": 0.4,
>	"life_stage": "pre_retirement"
>	}'

1	{
2	"results": [
3	{
4	"comment_id": "UgxAbc123",
5	"text": "I'm 54 and still have 11 years left on the mortgage...",
6	"like_count": 142,
7	"emotional_register": "anxiety_fear",
8	"life_stage": "pre_retirement",
9	"product_named": [
10	"mortgage"
11	],
12	"similarity": 0.87
13	}
14	]
15	}