Embed a search query
Returns the embedding vector for a search string, using the same model that indexed the recipes. Use it to run your own semantic search over recipes you have synced locally — embedding the query client-side with a different model would not produce comparable distances.
Returns the embedding vector for a search string, using the same model that indexed the recipes. Use it to run your own semantic search over recipes you have synced locally — embedding the query client-side with a different model would not produce comparable distances.
Authorization
bearerAuth A Clerk session JWT. Obtain one with a Clerk frontend SDK
(session.getToken()) or, for server-to-server use, with a Clerk
machine token. The token's sub claim is the Flambe user id that scopes
every request.
In: header
Request Body
application/json
TypeScript Definitions
Use the request body type in TypeScript.
Response Body
application/json
application/json
application/json
application/json
curl -X POST "https://example.com/api/ai/embed-query" \ -H "Content-Type: application/json" \ -d '{ "query": "something warm with lentils" }'{ "embedding": [ 0 ]}Create a realtime assistant session POST
Mints a short-lived client secret for the hands-free cooking assistant, so a client can open a realtime connection without ever holding a long-lived provider key. Scope it to a recipe with `recipeId`. The secret expires quickly — request one per session, at the moment the user starts talking, rather than caching it.
Log in (legacy) POST
Legacy username/password login, retained for older clients. Use Clerk instead.