adding

server.js

@@ -38,18 +38,18 @@ app.post("/api/generate/chat/completions", validateApiKey, async (req, res) => {
   try {
     const { model, messages } = req.body;
 
-    // 🧠 Convert OpenAI messages to plain prompt
-    const prompt = messages
-      .map(
-        (msg) =>
-          `${msg.role === "system" ? "" : msg.role + ": "}${msg.content}`,
-      )
-      .join("\n");
+    const prompt =
+      messages
+        .map((msg) => {
+          if (msg.role === "system") return `${msg.content}\n`;
+          return `${msg.role === "user" ? "User" : "Assistant"}: ${msg.content}`;
+        })
+        .join("\n") + "\nAssistant:";
 
     console.log("Generated prompt for Ollama:\n", prompt);
 
     const response = await axios.post(
-      "http://localhost:11434/api/generate/chat/completions",
+      "http://localhost:11434/api/generate",
       {
         model: model || "deepseek-r1:latest",
         prompt,
@@ -58,8 +58,24 @@ app.post("/api/generate/chat/completions", validateApiKey, async (req, res) => {
       { responseType: "stream" },
     );
 
+    // 🔥 Manual stream reading for logging
     console.log("Streaming response from Ollama...");
-    response.data.pipe(res);
+    response.data.on("data", (chunk) => {
+      const text = chunk.toString("utf8");
+      console.log("🧩 Chunk:", text);
+      res.write(text); // forward to client
+    });
+
+    response.data.on("end", () => {
+      console.log("✅ Streaming complete.");
+      res.end();
+    });
+
+    response.data.on("error", (err) => {
+      console.error("❌ Stream error:", err.message);
+      res.end();
+    });
+
   } catch (error) {
     console.error(
       "Error forwarding request to Ollama:",