mstar.api_server.openai.router#

FastAPI routes for the OpenAI-compatible API.

Endpoints stay model-agnostic: each looks up the loaded model’s adapter, checks the surface is supported, and delegates to a serving handler. The native /generate endpoint is unaffected.

Functions

async mstar.api_server.openai.router.audio_speech(request)[source]#
Parameters:

request (SpeechRequest)

async mstar.api_server.openai.router.chat_completions(request)[source]#
Parameters:

request (ChatCompletionRequest)

async mstar.api_server.openai.router.images_edits(request)[source]#
Parameters:

request (Request)

async mstar.api_server.openai.router.images_generations(request)[source]#
Parameters:

request (ImageGenerationRequest)

async mstar.api_server.openai.router.list_models()[source]#