Embeddings — /v1/embeddings

The embeddings endpoints wrap LlamaBridge embedding inference. Use this for RAG pipelines, semantic search, and similarity ranking.

Initialize

POST /v1/embeddings/init

Load an embedding model. Must be called before /embed.

Request:

{ "modelPath": "/models/nomic-embed-text-v1.5.Q8_0.gguf" }

Response:

{ "ok": true }

Returns 400 on failure.

Generate an embedding

POST /v1/embeddings/embed

Compute an embedding vector for the given input string.

Request:

{ "input": "Kotlin Multiplatform" }

Response:

{ "embedding": [0.12, -0.04, 0.87, "..."] }

The embedding array length matches the model’s output dimension. Typical values:

• nomic-embed-text-v1.5 → 768 dimensions
• mxbai-embed-large → 1024 dimensions

Notes

• One model at a time. Call init again to swap models.
• Embedding models are different from generation models — use a model specifically trained for embeddings.
• The server does not expose a /release endpoint for embeddings; the embedding model shares the LlamaBridge lifecycle with the generation model.