fix(onboard): use 127.0.0.1 instead of localhost for local inference detection

bin/lib/local-inference.js

@@ -21,9 +21,9 @@ function getLocalProviderBaseUrl(provider) {
 function getLocalProviderHealthCheck(provider) {
   switch (provider) {
     case "vllm-local":
-      return "curl -sf http://localhost:8000/v1/models 2>/dev/null";
+      return "curl -sf http://127.0.0.1:8000/v1/models 2>/dev/null";
     case "ollama-local":
-      return "curl -sf http://localhost:11434/api/tags 2>/dev/null";
+      return "curl -sf http://127.0.0.1:11434/api/tags 2>/dev/null";
     default:
       return null;
   }
@@ -52,12 +52,12 @@ function validateLocalProvider(provider, runCapture) {
       case "vllm-local":
         return {
           ok: false,
-          message: "Local vLLM was selected, but nothing is responding on http://localhost:8000.",
+          message: "Local vLLM was selected, but nothing is responding on http://127.0.0.1:8000.",
         };
       case "ollama-local":
         return {
           ok: false,
-          message: "Local Ollama was selected, but nothing is responding on http://localhost:11434.",
+          message: "Local Ollama was selected, but nothing is responding on http://127.0.0.1:11434.",
         };
       default:
         return { ok: false, message: "The selected local inference provider is unavailable." };
@@ -79,13 +79,13 @@ function validateLocalProvider(provider, runCapture) {
       return {
         ok: false,
         message:
-          "Local vLLM is responding on localhost, but containers cannot reach http://host.openshell.internal:8000. Ensure the server is reachable from containers, not only from the host shell.",
+          "Local vLLM is responding on 127.0.0.1, but containers cannot reach http://host.openshell.internal:8000. Ensure the server is reachable from containers, not only from the host shell.",
       };
     case "ollama-local":
       return {
         ok: false,
         message:
-          "Local Ollama is responding on localhost, but containers cannot reach http://host.openshell.internal:11434. Ensure Ollama listens on 0.0.0.0:11434 instead of 127.0.0.1 so sandboxes can reach it.",
+          "Local Ollama is responding on 127.0.0.1, but containers cannot reach http://host.openshell.internal:11434. Ensure Ollama listens on 0.0.0.0:11434 instead of 127.0.0.1 so sandboxes can reach it.",
       };
     default:
       return { ok: false, message: "The selected local inference provider is unavailable from containers." };
@@ -123,7 +123,7 @@ function getOllamaWarmupCommand(model, keepAlive = "15m") {
     stream: false,
     keep_alive: keepAlive,
   });
-  return `nohup curl -s http://localhost:11434/api/generate -H 'Content-Type: application/json' -d ${shellQuote(payload)} >/dev/null 2>&1 &`;
+  return `nohup curl -s http://127.0.0.1:11434/api/generate -H 'Content-Type: application/json' -d ${shellQuote(payload)} >/dev/null 2>&1 &`;
 }
 
 function getOllamaProbeCommand(model, timeoutSeconds = 120, keepAlive = "15m") {
@@ -133,7 +133,7 @@ function getOllamaProbeCommand(model, timeoutSeconds = 120, keepAlive = "15m") {
     stream: false,
     keep_alive: keepAlive,
   });
-  return `curl -sS --max-time ${timeoutSeconds} http://localhost:11434/api/generate -H 'Content-Type: application/json' -d ${shellQuote(payload)} 2>/dev/null`;
+  return `curl -sS --max-time ${timeoutSeconds} http://127.0.0.1:11434/api/generate -H 'Content-Type: application/json' -d ${shellQuote(payload)} 2>/dev/null`;
 }
 
 function validateOllamaModel(model, runCapture) {

bin/lib/nim.js

@@ -152,7 +152,7 @@ function waitForNimHealth(port = 8000, timeout = 300) {
 
   while ((Date.now() - start) / 1000 < timeout) {
     try {
-      const result = runCapture(`curl -sf http://localhost:${safePort}/v1/models`, {
+      const result = runCapture(`curl -sf http://127.0.0.1:${safePort}/v1/models`, {
         ignoreError: true,
       });
       if (result) {
@@ -186,7 +186,7 @@ function nimStatus(sandboxName) {
 
     let healthy = false;
     if (state === "running") {
-      const health = runCapture(`curl -sf http://localhost:8000/v1/models 2>/dev/null`, {
+      const health = runCapture(`curl -sf http://127.0.0.1:8000/v1/models 2>/dev/null`, {
         ignoreError: true,
       });
       healthy = !!health;

bin/lib/onboard.js

@@ -527,8 +527,8 @@ async function setupNim(sandboxName, gpu) {
 
   // Detect local inference options
   const hasOllama = !!runCapture("command -v ollama", { ignoreError: true });
-  const ollamaRunning = !!runCapture("curl -sf http://localhost:11434/api/tags 2>/dev/null", { ignoreError: true });
-  const vllmRunning = !!runCapture("curl -sf http://localhost:8000/v1/models 2>/dev/null", { ignoreError: true });
+  const ollamaRunning = !!runCapture("curl -sf http://127.0.0.1:11434/api/tags 2>/dev/null", { ignoreError: true });
+  const vllmRunning = !!runCapture("curl -sf http://127.0.0.1:8000/v1/models 2>/dev/null", { ignoreError: true });
   const requestedProvider = isNonInteractive() ? getNonInteractiveProvider() : null;
   const requestedModel = isNonInteractive() ? getNonInteractiveModel(requestedProvider || "cloud") : null;
   // Build options list — only show local options with NEMOCLAW_EXPERIMENTAL=1
@@ -546,14 +546,14 @@ async function setupNim(sandboxName, gpu) {
     options.push({
       key: "ollama",
       label:
-        `Local Ollama (localhost:11434)${ollamaRunning ? " — running" : ""}` +
+        `Local Ollama (127.0.0.1:11434)${ollamaRunning ? " — running" : ""}` +
         (ollamaRunning ? " (suggested)" : ""),
     });
   }
   if (EXPERIMENTAL && vllmRunning) {
     options.push({
       key: "vllm",
-      label: "Existing vLLM instance (localhost:8000) — running [experimental] (suggested)",
+      label: "Existing vLLM instance (127.0.0.1:8000) — running [experimental] (suggested)",
     });
   }
 
@@ -649,7 +649,7 @@ async function setupNim(sandboxName, gpu) {
         run("OLLAMA_HOST=0.0.0.0:11434 ollama serve > /dev/null 2>&1 &", { ignoreError: true });
         sleep(2);
       }
-      console.log("  ✓ Using Ollama on localhost:11434");
+      console.log("  ✓ Using Ollama on 127.0.0.1:11434");
       provider = "ollama-local";
       if (isNonInteractive()) {
         model = requestedModel || getDefaultOllamaModel(runCapture);
@@ -662,15 +662,15 @@ async function setupNim(sandboxName, gpu) {
       console.log("  Starting Ollama...");
       run("OLLAMA_HOST=0.0.0.0:11434 ollama serve > /dev/null 2>&1 &", { ignoreError: true });
         sleep(2);
-      console.log("  ✓ Using Ollama on localhost:11434");
+      console.log("  ✓ Using Ollama on 127.0.0.1:11434");
       provider = "ollama-local";
       if (isNonInteractive()) {
         model = requestedModel || getDefaultOllamaModel(runCapture);
       } else {
         model = await promptOllamaModel();
       }
     } else if (selected.key === "vllm") {
-      console.log("  ✓ Using existing vLLM on localhost:8000");
+      console.log("  ✓ Using existing vLLM on 127.0.0.1:8000");
       provider = "vllm-local";
       model = "vllm-local";
     }

test/local-inference.test.js

@@ -37,7 +37,7 @@ describe("local inference helpers", () => {
   it("returns the expected health check command for ollama-local", () => {
     assert.equal(
       getLocalProviderHealthCheck("ollama-local"),
-      "curl -sf http://localhost:11434/api/tags 2>/dev/null",
+      "curl -sf http://127.0.0.1:11434/api/tags 2>/dev/null",
     );
   });
 
@@ -61,7 +61,7 @@ describe("local inference helpers", () => {
   it("returns a clear error when ollama-local is unavailable", () => {
     const result = validateLocalProvider("ollama-local", () => "");
     assert.equal(result.ok, false);
-    assert.match(result.message, /http:\/\/localhost:11434/);
+    assert.match(result.message, /http:\/\/127\.0\.0\.1:11434/);
   });
 
   it("returns a clear error when ollama-local is not reachable from containers", () => {
@@ -78,7 +78,7 @@ describe("local inference helpers", () => {
   it("returns a clear error when vllm-local is unavailable", () => {
     const result = validateLocalProvider("vllm-local", () => "");
     assert.equal(result.ok, false);
-    assert.match(result.message, /http:\/\/localhost:8000/);
+    assert.match(result.message, /http:\/\/127\.0\.0\.1:8000/);
   });
 
   it("parses model names from ollama list output", () => {
@@ -121,14 +121,14 @@ describe("local inference helpers", () => {
 
   it("builds a background warmup command for ollama models", () => {
     const command = getOllamaWarmupCommand("nemotron-3-nano:30b");
-    assert.match(command, /^nohup curl -s http:\/\/localhost:11434\/api\/generate /);
+    assert.match(command, /^nohup curl -s http:\/\/127\.0\.0\.1:11434\/api\/generate /);
     assert.match(command, /"model":"nemotron-3-nano:30b"/);
     assert.match(command, /"keep_alive":"15m"/);
   });
 
   it("builds a foreground probe command for ollama models", () => {
     const command = getOllamaProbeCommand("nemotron-3-nano:30b");
-    assert.match(command, /^curl -sS --max-time 120 http:\/\/localhost:11434\/api\/generate /);
+    assert.match(command, /^curl -sS --max-time 120 http:\/\/127\.0\.0\.1:11434\/api\/generate /);
     assert.match(command, /"model":"nemotron-3-nano:30b"/);
   });
 

test/onboard-selection.test.js

@@ -34,9 +34,9 @@ credentials.prompt = async (message) => {
 credentials.ensureApiKey = async () => {};
 runner.runCapture = (command) => {
   if (command.includes("command -v ollama")) return "/usr/bin/ollama";
-  if (command.includes("localhost:11434/api/tags")) return JSON.stringify({ models: [{ name: "nemotron-3-nano:30b" }] });
+  if (command.includes("127.0.0.1:11434/api/tags")) return JSON.stringify({ models: [{ name: "nemotron-3-nano:30b" }] });
   if (command.includes("ollama list")) return "nemotron-3-nano:30b  abc  24 GB  now\\nqwen3:32b  def  20 GB  now";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   return "";
 };
 registry.updateSandbox = (_name, update) => updates.push(update);