82 lines
2.6 KiB
Python
82 lines
2.6 KiB
Python
|
"""
|
||
|
|
token_server.py
|
||
|
|
Serves:
|
||
|
|
GET /token?room=<name> → LiveKit JWT token
|
||
|
|
GET /models → list of Ollama/LM Studio models
|
||
|
|
static/index.html → browser UI
|
||
|
|
"""
|
||
|
|
|
||
|
|
import os, json, asyncio, httpx
|
||
|
|
from pathlib import Path
|
||
|
|
from dotenv import load_dotenv
|
||
|
|
from fastapi import FastAPI, Query
|
||
|
|
from fastapi.responses import JSONResponse
|
||
|
|
from fastapi.staticfiles import StaticFiles
|
||
|
|
import uvicorn
|
||
|
|
from livekit.api import AccessToken, VideoGrants, RoomConfiguration, RoomAgentDispatch
|
||
|
|
|
||
|
|
load_dotenv()
|
||
|
|
|
||
|
|
LIVEKIT_API_KEY = os.getenv("LIVEKIT_API_KEY", "devkey")
|
||
|
|
LIVEKIT_API_SECRET = os.getenv("LIVEKIT_API_SECRET", "secret")
|
||
|
|
OLLAMA_BASE_URL = os.getenv("OLLAMA_BASE_URL", "http://localhost:11434/v1")
|
||
|
|
|
||
|
|
app = FastAPI()
|
||
|
|
|
||
|
|
@app.get("/token")
|
||
|
|
async def get_token(
|
||
|
|
room: str = Query("my-room"),
|
||
|
|
identity: str = Query("browser-user"),
|
||
|
|
model: str = Query(""),
|
||
|
|
prompt: str = Query(""),
|
||
|
|
):
|
||
|
|
attributes = {}
|
||
|
|
if model:
|
||
|
|
attributes["model"] = model
|
||
|
|
if prompt:
|
||
|
|
attributes["prompt"] = prompt
|
||
|
|
|
||
|
|
token = (
|
||
|
|
AccessToken(LIVEKIT_API_KEY, LIVEKIT_API_SECRET)
|
||
|
|
.with_identity(identity)
|
||
|
|
.with_name(identity)
|
||
|
|
.with_grants(VideoGrants(room_join=True, room=room))
|
||
|
|
.with_room_config(RoomConfiguration(
|
||
|
|
agents=[RoomAgentDispatch(agent_name="")]
|
||
|
|
))
|
||
|
|
)
|
||
|
|
if attributes:
|
||
|
|
token = token.with_attributes(attributes)
|
||
|
|
|
||
|
|
return {"token": token.to_jwt(), "room": room}
|
||
|
|
|
||
|
|
@app.get("/models")
|
||
|
|
async def list_models():
|
||
|
|
"""Proxy to Ollama/LM Studio model list endpoint."""
|
||
|
|
try:
|
||
|
|
async with httpx.AsyncClient(timeout=5) as client:
|
||
|
|
# OpenAI-compat endpoint (works for Ollama & LM Studio)
|
||
|
|
r = await client.get(OLLAMA_BASE_URL.rstrip("/v1").rstrip("/") + "/api/tags")
|
||
|
|
if r.status_code == 200:
|
||
|
|
data = r.json()
|
||
|
|
models = [m["name"] for m in data.get("models", [])]
|
||
|
|
return {"models": models}
|
||
|
|
except Exception:
|
||
|
|
pass
|
||
|
|
# Fallback: OpenAI /v1/models
|
||
|
|
try:
|
||
|
|
async with httpx.AsyncClient(timeout=5) as client:
|
||
|
|
r = await client.get(OLLAMA_BASE_URL.rstrip("/") + "/models")
|
||
|
|
if r.status_code == 200:
|
||
|
|
data = r.json()
|
||
|
|
models = [m["id"] for m in data.get("data", [])]
|
||
|
|
return {"models": models}
|
||
|
|
except Exception:
|
||
|
|
pass
|
||
|
|
return {"models": [], "error": "Could not reach local model server"}
|
||
|
|
|
||
|
|
app.mount("/", StaticFiles(directory="static", html=True), name="static")
|
||
|
|
|
||
|
|
if __name__ == "__main__":
|
||
|
|
uvicorn.run(app, host="0.0.0.0", port=7731)
|