Allow overriding the model with env var

app/agent.py

@@ -1,5 +1,6 @@
 import asyncio
 import logging
+import os
 from collections.abc import AsyncIterable
 
 from haystack.dataclasses import ChatMessage
@@ -14,6 +15,8 @@ from app.queue import StreamingQueue
 
 logger = logging.getLogger(__name__)
 
+MODEL = os.environ.get("MODEL", "nvidia/nemotron-3-super-120b-a12b")
+
 _QUERY_ONLY_SCHEMA = {
     "type": "object",
     "properties": {
@@ -28,8 +31,7 @@ _QUERY_ONLY_SCHEMA = {
 
 def make_generator() -> NvidiaChatGenerator:
     return NvidiaChatGenerator(
-        # model="openai/gpt-oss-120b",
-        model="nvidia/nemotron-3-super-120b-a12b",
+        model=MODEL,
         generation_kwargs={
             "stream": True,
             "max_tokens": 96_000,

app/api.py

@@ -6,7 +6,7 @@ from fastapi import FastAPI
 from fastapi.responses import FileResponse
 from fastapi.staticfiles import StaticFiles
 
-from app.agent import run_agent
+from app.agent import MODEL, run_agent
 from app.models import AgentRequest, TextChunk
 
 logging.basicConfig(
@@ -22,6 +22,11 @@ _STATIC = Path(__file__).parent / "static"
 app.mount("/static", StaticFiles(directory=_STATIC), name="static")
 
 
+@app.get("/config")
+async def config():
+    return {"model": MODEL}
+
+
 @app.get("/")
 async def index() -> FileResponse:
     return FileResponse(_STATIC / "index.html")

app/static/index.html

@@ -41,7 +41,7 @@
         the agent searches our Google Drive with conference materials and gets you up to speed in seconds.
       </p>
       <p class="hero-powered">
-        Powered by <code class="hero-model">nvidia/nemotron-3-super-120b-a12b</code>
+        Powered by <code class="hero-model">…</code>
       </p>
     </div>
 
@@ -137,6 +137,10 @@
   </div>
 
   <script>
+    fetch('/config').then(r => r.json()).then(cfg => {
+      document.querySelector('.hero-model').textContent = cfg.model;
+    });
+
     marked.use({ breaks: true });
 
     // ── DOM ───────────────────────────────────────────────────────────