Update app.py

app.py

@@ -165,6 +165,151 @@ async def chat(request: Request, authorization: str = Header(None)):
         }
 
     return JSONResponse(data, headers={"Access-Control-Allow-Origin": "*"})
+
+
+
+
+
+
+
+# ---------------------------------------------------------------------
+# /v2/chat/completions  →  Puter.js → OpenAI compatible
+# ---------------------------------------------------------------------
+@api.post("/v2/chat/completions")
+async def puter_chat(request: Request, authorization: str = Header(None)):
+    check_key(authorization)
+
+    payload = await request.json()
+    model       = payload.get("model")
+    messages    = payload.get("messages", [])
+    temperature = payload.get("temperature", 1.0)
+    max_tokens  = payload.get("max_tokens")
+    stream      = payload.get("stream", False)
+
+    # Convert OpenAI-style messages → single string prompt for Puter.js
+    prompt = "\n".join([f"{m['role']}: {m['content']}" for m in messages])
+
+    # Node helper input
+    node_payload = json.dumps({
+        "prompt": prompt,
+        "model": model,
+        "temperature": temperature,
+        "max_tokens": max_tokens,
+        "stream": False     # streaming handled later
+    })
+
+    # ------------------------------------------------------------------
+    # Non-streaming
+    # ------------------------------------------------------------------
+    if not stream:
+        proc = subprocess.Popen(
+            ["node", "puter_helper.js"],
+            stdin=subprocess.PIPE,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True
+        )
+
+        stdout, stderr = proc.communicate(node_payload)
+
+        if stderr:
+            raise HTTPException(status_code=500, detail=f"Node error: {stderr}")
+
+        try:
+            node_out = json.loads(stdout)
+        except:
+            raise HTTPException(status_code=500, detail=f"Bad Node output: {stdout}")
+
+        if not node_out.get("ok"):
+            raise HTTPException(status_code=502, detail=node_out.get("error"))
+
+        final_text = node_out["result"]
+
+        return {
+            "id": "chatcmpl-puter",
+            "object": "chat.completion",
+            "model": model,
+            "choices": [
+                {
+                    "index": 0,
+                    "message": {
+                        "role": "assistant",
+                        "content": final_text
+                    },
+                    "finish_reason": "stop"
+                }
+            ]
+        }
+
+    # ------------------------------------------------------------------
+    # Streaming path: /v2/chat/completions?stream=true
+    # ------------------------------------------------------------------
+    async def stream_generator():
+        # Because Puter.js Node helper is not streaming yet,
+        # we emulate SSE streaming by splitting text gradually.
+        proc = subprocess.Popen(
+            ["node", "puter_helper.js"],
+            stdin=subprocess.PIPE,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True
+        )
+
+        stdout, stderr = proc.communicate(node_payload)
+
+        if stderr:
+            yield f"data: {{\"error\": \"{stderr}\"}}\n\n"
+            yield "data: [DONE]\n\n"
+            return
+
+        try:
+            node_out = json.loads(stdout)
+        except:
+            yield f"data: {{\"error\": \"Bad Node output\"}}\n\n"
+            yield "data: [DONE]\n\n"
+            return
+
+        if not node_out.get("ok"):
+            yield f"data: {{\"error\": \"{node_out.get('error')}\"}}\n\n"
+            yield "data: [DONE]\n\n"
+            return
+
+        full_text = node_out["result"]
+
+        # Send word-by-word as streaming chunks
+        for word in full_text.split():
+            chunk = {
+                "id": "chatcmpl-puter-stream",
+                "object": "chat.completion.chunk",
+                "model": model,
+                "choices": [
+                    {
+                        "index": 0,
+                        "delta": {"content": word + " "},
+                        "finish_reason": None,
+                    }
+                ]
+            }
+            yield f"data: {json.dumps(chunk)}\n\n"
+            await asyncio.sleep(0.02)
+
+        yield "data: [DONE]\n\n"
+
+    return StreamingResponse(
+        stream_generator(),
+        media_type="text/event-stream",
+        headers={"Access-Control-Allow-Origin": "*"},
+    )
+
+
+
+
+
+
+
+
+
+    
 # ---------------------------------------------------------------------
 # Minimal Gradio UI (to make HF Space start)
 # ---------------------------------------------------------------------