open-agents-ai 0.185.91 → 0.185.92

package/README.md

@@ -646,6 +646,37 @@ curl -s http://localhost:11435/v1/chat \
 # Response: {"session_id": "abc123", "message": {"role": "assistant", "content": "..."}}
 ```
 
+**Request body:**
+```json
+{
+  "message": "What is happening in the world?",
+  "model": "qwen3.5:9b",
+  "session_id": "optional-uuid-from-previous-response",
+  "stream": true,
+  "max_tokens": 4096
+}
+```
+
+**Response (non-streaming):**
+```json
+{
+  "session_id": "abc123-def4-5678-ghij-klmnopqrstuv",
+  "message": {
+    "role": "assistant",
+    "content": "Here are the major events happening today..."
+  }
+}
+```
+
+**Response (streaming `stream: true`):** Server-Sent Events:
+```
+data: {"type":"tool_call","tool":"web_search","args":{"query":"world news today"}}
+data: {"type":"tool_result","output":"Top results: ..."}
+data: {"id":"chatcmpl-abc","object":"chat.completion.chunk","choices":[{"delta":{"content":"Based on..."}}]}
+data: {"type":"complete","turns":"3","tokens":"12,450","duration":8500}
+data: [DONE]
+```
+
 **Session management:** Each chat message returns a `session_id`. Send it back to maintain conversation context across turns:
 
 ```bash

package/dist/index.js

@@ -65552,7 +65552,9 @@ body {
   <button class="tab" onclick="switchTab('jobs')" id="tab-jobs" style="background:none;border:none;border-bottom:2px solid transparent;color:#555;padding:6px 16px;font-family:inherit;font-size:0.7rem;cursor:pointer">dashboard</button>
   <button class="tab" onclick="switchTab('config')" id="tab-config" style="background:none;border:none;border-bottom:2px solid transparent;color:#555;padding:6px 16px;font-family:inherit;font-size:0.7rem;cursor:pointer">config</button>
   <button class="tab" onclick="switchTab('activity')" id="tab-activity" style="background:none;border:none;border-bottom:2px solid transparent;color:#555;padding:6px 16px;font-family:inherit;font-size:0.7rem;cursor:pointer">activity</button>
-  <span id="token-counter" style="margin-left:auto;font-size:0.6rem;color:#555">0 tokens</span>
+  <span id="sys-metrics" style="margin-left:auto;font-size:0.6rem;color:#555"></span>
+  <span id="update-btn" style="display:none;background:#3a2a10;border:1px solid #b2920a;color:#b2920a;padding:2px 8px;border-radius:3px;font-family:inherit;font-size:0.6rem;cursor:pointer" onclick="doUpdate()">update</span>
+  <span id="token-counter" style="font-size:0.6rem;color:#555">0 tokens</span>
 </div>
 <div style="display:flex;flex:1;overflow:hidden">
 <div id="workspace-sidebar" style="display:none;width:250px;background:#1e1e22;border-right:1px solid #2a2a30;overflow-y:auto;padding:8px;flex-shrink:0;font-size:0.7rem">
@@ -66243,10 +66245,44 @@ function toggleSandbox() {
   btn.style.borderColor = sandboxMode === 'none' ? '#3a3a42' : '#b2920a';
 }
 
+// Live system metrics polling
+async function pollMetrics() {
+  try {
+    const r = await fetch('/v1/system', { headers: headers() });
+    const d = await r.json();
+    const el = document.getElementById('sys-metrics');
+    const parts = [];
+    if (d.cpu_pct !== undefined) parts.push('CPU ' + d.cpu_pct + '%');
+    if (d.ram_used_pct !== undefined) parts.push('RAM ' + d.ram_used_pct + '%');
+    if (d.gpu_utilization?.length) {
+      const g = d.gpu_utilization[0];
+      parts.push('GPU ' + g.gpu_pct + '%');
+      parts.push('VRAM ' + g.vram_used_gb + '/' + g.vram_total_gb + 'GB');
+    }
+    if (el) el.textContent = parts.join(' | ');
+    // Update button
+    const btn = document.getElementById('update-btn');
+    if (btn && d.update_available && d.latest_version) {
+      btn.style.display = 'inline-block';
+      btn.textContent = 'update v' + d.latest_version;
+      btn.title = 'Current: v' + d.version + ' \u2192 Latest: v' + d.latest_version;
+    }
+  } catch {}
+}
+
+function doUpdate() {
+  if (confirm('Update Open Agents to latest version? This will restart the server.')) {
+    fetch('/v1/commands/update', { method: 'POST', headers: headers() }).catch(() => {});
+    document.getElementById('update-btn').textContent = 'updating...';
+  }
+}
+
 // Init
 checkHealth();
 loadModels();
+pollMetrics();
 setInterval(checkHealth, 30000);
+setInterval(pollMetrics, 10000);
 input.focus();
 </script>
 </body>
@@ -67946,6 +67982,7 @@ async function handleRequest(req, res, ollamaUrl, verbose) {
     }
     if (pathname === "/v1/system" && method === "GET") {
       const os = __require("node:os");
+      const version = getVersion3();
       const { execSync: es } = __require("node:child_process");
       let gpus = [];
       try {
@@ -67958,15 +67995,43 @@ async function handleRequest(req, res, ollamaUrl, verbose) {
       } catch {
       }
       const totalVram = gpus.reduce((s, g) => s + g.vram_gb, 0);
+      const freeMem = os.freemem();
+      const totalMem = os.totalmem();
+      const ramUsedPct = Math.round((1 - freeMem / totalMem) * 100);
+      const cpuLoad = os.loadavg()[0] ?? 0;
+      const cpuPct = Math.min(100, Math.round(cpuLoad / os.cpus().length * 100));
+      let gpuUtil = [];
+      try {
+        const util = es("nvidia-smi --query-gpu=utilization.gpu,memory.used,memory.total --format=csv,noheader,nounits", { encoding: "utf8", timeout: 3e3, stdio: "pipe" });
+        for (const line of util.trim().split("\n")) {
+          const [pct, used, total] = line.split(",").map((s) => parseInt(s.trim(), 10));
+          if (!isNaN(pct) && !isNaN(used) && !isNaN(total)) {
+            gpuUtil.push({ gpu_pct: pct, vram_used_gb: Math.round(used / 1024 * 10) / 10, vram_total_gb: Math.round(total / 1024) });
+          }
+        }
+      } catch {
+      }
+      let latestVersion = null;
+      try {
+        const ver = es("npm view open-agents-ai version 2>/dev/null", { encoding: "utf8", timeout: 5e3, stdio: "pipe" }).trim();
+        if (ver && ver !== version)
+          latestVersion = ver;
+      } catch {
+      }
       jsonResponse(res, 200, {
         gpu: gpus,
+        gpu_utilization: gpuUtil,
         total_vram_gb: totalVram,
-        ram_gb: Math.round(os.totalmem() / 1024 ** 3),
+        ram_gb: Math.round(totalMem / 1024 ** 3),
+        ram_used_pct: ramUsedPct,
         cpu: os.cpus()[0]?.model ?? "unknown",
+        cpu_pct: cpuPct,
         cores: os.cpus().length,
         platform: process.platform,
         node: process.version,
-        // Model compatibility recommendations
+        version,
+        latest_version: latestVersion,
+        update_available: !!latestVersion,
         recommended_max_params: totalVram >= 80 ? "120B+" : totalVram >= 48 ? "70B" : totalVram >= 24 ? "27B" : totalVram >= 8 ? "9B" : "4B"
       });
       return;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "open-agents-ai",
-  "version": "0.185.91",
+  "version": "0.185.92",
   "description": "AI coding agent powered by open-source models (Ollama/vLLM) — interactive TUI with agentic tool-calling loop",
   "type": "module",
   "main": "./dist/index.js",