Skip to main content

Crate vera2_serve

Crate vera2_serve 

Source
Expand description

Vera HTTP API server.

Exposes OpenAI-compatible inference endpoints for standard vera clients:

POST /v1/embeddings   OpenAI format  (EMBEDDING_MODEL_BASE_URL)
POST /v1/rerank       Cohere/Jina format  (RERANKER_MODEL_BASE_URL)
GET  /v1/health       liveness + model info

A regular vera client configured with vera setup --api pointing at http://host:port/v1 will work without any modifications.

Modules§

types
Request and response types for the Vera HTTP API.

Structs§

AppState
Shared state injected into every handler.

Enums§

CacheMode
How long a loaded model stays resident.

Functions§

run_server
Start the Vera HTTP server.