Embeddings
POST /v1/embeddings through Maple Proxy. Create vector embeddings for search and retrieval with OpenAI's request format and a Maple embedding model.
POST /v1/embeddingsCreates embeddings for search and retrieval. Send the same request body you would send to OpenAI’s Embeddings API.
curl http://127.0.0.1:8080/v1/embeddings \
-H "Authorization: Bearer $MAPLE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "nomic-embed-text",
"input": "Generate an embedding for this text"
}'What differs from OpenAI
- Models. Use
nomic-embed-text, Maple’s embedding model, not an OpenAI one. It is also the default if you leavemodelout. Embeddings are billed on input tokens only; see models. - Pass-through body. The proxy forwards the body exactly as sent and returns the response unchanged. Options the model doesn’t support are handled by Maple’s backend, not the proxy.
- Limits and errors. The same 50 MiB request limit, timeouts and error statuses apply as for chat completions.
Last updated