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

src/lib/local-inference.test.ts

@@ -47,25 +47,25 @@ describe("local inference helpers", () => {
   });
 
   it("returns the expected validation URL for vllm-local", () => {
-    expect(getLocalProviderValidationBaseUrl("vllm-local")).toBe("http://localhost:8000/v1");
+    expect(getLocalProviderValidationBaseUrl("vllm-local")).toBe("http://127.0.0.1:8000/v1");
   });
 
   it("returns the expected health check command for ollama-local", () => {
     expect(getLocalProviderHealthEndpoint("ollama-local")).toBe(
-      "http://localhost:11434/api/tags",
+      "http://127.0.0.1:11434/api/tags",
     );
     expect(getLocalProviderLabel("ollama-local")).toBe("Local Ollama");
     expect(getLocalProviderHealthCheck("ollama-local")).toBe(
-      "curl -sf http://localhost:11434/api/tags 2>/dev/null",
+      "curl -sf http://127.0.0.1:11434/api/tags 2>/dev/null",
     );
   });
 
   it("returns the expected validation and health check commands for vllm-local", () => {
-    expect(getLocalProviderValidationBaseUrl("ollama-local")).toBe("http://localhost:11434/v1");
-    expect(getLocalProviderHealthEndpoint("vllm-local")).toBe("http://localhost:8000/v1/models");
+    expect(getLocalProviderValidationBaseUrl("ollama-local")).toBe("http://127.0.0.1:11434/v1");
+    expect(getLocalProviderHealthEndpoint("vllm-local")).toBe("http://127.0.0.1:8000/v1/models");
     expect(getLocalProviderLabel("vllm-local")).toBe("Local vLLM");
     expect(getLocalProviderHealthCheck("vllm-local")).toBe(
-      "curl -sf http://localhost:8000/v1/models 2>/dev/null",
+      "curl -sf http://127.0.0.1:8000/v1/models 2>/dev/null",
     );
     expect(getLocalProviderContainerReachabilityCheck("vllm-local")).toBe(
       `docker run --rm --add-host host.openshell.internal:host-gateway ${CONTAINER_REACHABILITY_IMAGE} -sf http://host.openshell.internal:8000/v1/models 2>/dev/null`,
@@ -91,7 +91,7 @@ describe("local inference helpers", () => {
   it("returns a clear error when ollama-local is unavailable", () => {
     const result = validateLocalProvider("ollama-local", () => "");
     expect(result.ok).toBe(false);
-    expect(result.message).toMatch(/http:\/\/localhost:11434/);
+    expect(result.message).toMatch(/http:\/\/127.0.0.1:11434/);
   });
 
   it("returns a clear error when ollama-local is not reachable from containers", () => {
@@ -108,7 +108,7 @@ describe("local inference helpers", () => {
   it("returns a clear error when vllm-local is unavailable", () => {
     const result = validateLocalProvider("vllm-local", () => "");
     expect(result.ok).toBe(false);
-    expect(result.message).toMatch(/http:\/\/localhost:8000/);
+    expect(result.message).toMatch(/http:\/\/127.0.0.1:8000/);
   });
 
   it("probes local provider health successfully", () => {
@@ -126,8 +126,8 @@ describe("local inference helpers", () => {
     expect(result).toEqual({
       ok: true,
       providerLabel: "Local Ollama",
-      endpoint: "http://localhost:11434/api/tags",
-      detail: "Local Ollama is reachable on http://localhost:11434/api/tags.",
+      endpoint: "http://127.0.0.1:11434/api/tags",
+      detail: "Local Ollama is reachable on http://127.0.0.1:11434/api/tags.",
     });
   });
 
@@ -146,7 +146,7 @@ describe("local inference helpers", () => {
     expect(result?.ok).toBe(false);
     expect(result?.detail).toContain("Local Ollama is selected for inference");
     expect(result?.detail).toContain("Start Ollama and retry");
-    expect(result?.detail).toContain("http://localhost:11434/api/tags");
+    expect(result?.detail).toContain("http://127.0.0.1:11434/api/tags");
   });
 
   it("returns null when provider health probing is not supported", () => {
@@ -263,7 +263,7 @@ describe("local inference helpers", () => {
 
   it("builds a background warmup command for ollama models", () => {
     const command = getOllamaWarmupCommand("nemotron-3-nano:30b");
-    expect(command).toMatch(/^nohup curl -s http:\/\/localhost:11434\/api\/generate /);
+    expect(command).toMatch(/^nohup curl -s http:\/\/127.0.0.1:11434\/api\/generate /);
     expect(command).toMatch(/"model":"nemotron-3-nano:30b"/);
     expect(command).toMatch(/"keep_alive":"15m"/);
   });
@@ -276,7 +276,7 @@ describe("local inference helpers", () => {
 
   it("builds a foreground probe command for ollama models", () => {
     const command = getOllamaProbeCommand("nemotron-3-nano:30b");
-    expect(command).toMatch(/^curl -sS --max-time 120 http:\/\/localhost:11434\/api\/generate /);
+    expect(command).toMatch(/^curl -sS --max-time 120 http:\/\/127.0.0.1:11434\/api\/generate /);
     expect(command).toMatch(/"model":"nemotron-3-nano:30b"/);
   });
 

src/lib/local-inference.ts

@@ -54,9 +54,9 @@ export function getLocalProviderBaseUrl(provider: string): string | null {
 export function getLocalProviderValidationBaseUrl(provider: string): string | null {
   switch (provider) {
     case "vllm-local":
-      return "http://localhost:8000/v1";
+      return "http://127.0.0.1:8000/v1";
     case "ollama-local":
-      return "http://localhost:11434/v1";
+      return "http://127.0.0.1:11434/v1";
     default:
       return null;
   }
@@ -65,9 +65,9 @@ export function getLocalProviderValidationBaseUrl(provider: string): string | nu
 export function getLocalProviderHealthEndpoint(provider: string): string | null {
   switch (provider) {
     case "vllm-local":
-      return "http://localhost:8000/v1/models";
+      return "http://127.0.0.1:8000/v1/models";
     case "ollama-local":
-      return "http://localhost:11434/api/tags";
+      return "http://127.0.0.1:11434/api/tags";
     default:
       return null;
   }
@@ -170,13 +170,13 @@ export function validateLocalProvider(
       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.",
+            "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." };
@@ -198,13 +198,13 @@ export function validateLocalProvider(
       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 {
@@ -236,7 +236,7 @@ export function parseOllamaTags(output: unknown): string[] {
 }
 
 export function getOllamaModelOptions(runCapture: RunCaptureFn): string[] {
-  const tagsOutput = runCapture("curl -sf http://localhost:11434/api/tags 2>/dev/null", {
+  const tagsOutput = runCapture("curl -sf http://127.0.0.1:11434/api/tags 2>/dev/null", {
     ignoreError: true,
   });
   const tagsParsed = parseOllamaTags(tagsOutput);
@@ -275,7 +275,7 @@ export function getOllamaWarmupCommand(model: string, keepAlive = "15m"): string
     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 &`;
 }
 
 export function getOllamaProbeCommand(
@@ -289,7 +289,7 @@ export function getOllamaProbeCommand(
     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`;
 }
 
 export function validateOllamaModel(

src/lib/onboard.ts

@@ -28,6 +28,7 @@ const {
   getDefaultOllamaModel,
   getBootstrapOllamaModelOptions,
   getLocalProviderBaseUrl,
+  getLocalProviderHealthCheck,
   getLocalProviderValidationBaseUrl,
   getOllamaModelOptions,
   getOllamaWarmupCommand,
@@ -2823,10 +2824,10 @@ async function setupNim(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", {
+  const ollamaRunning = !!runCapture(getLocalProviderHealthCheck("ollama-local"), {
     ignoreError: true,
   });
-  const vllmRunning = !!runCapture("curl -sf http://localhost:8000/v1/models 2>/dev/null", {
+  const vllmRunning = !!runCapture(getLocalProviderHealthCheck("vllm-local"), {
     ignoreError: true,
   });
   const requestedProvider = isNonInteractive() ? getNonInteractiveProvider() : null;
@@ -3352,7 +3353,7 @@ async function setupNim(gpu) {
         credentialEnv = "OPENAI_API_KEY";
         endpointUrl = getLocalProviderBaseUrl(provider);
         // Query vLLM for the actual model ID
-        const vllmModelsRaw = runCapture("curl -sf http://localhost:8000/v1/models 2>/dev/null", {
+        const vllmModelsRaw = runCapture(getLocalProviderHealthCheck("vllm-local"), {
           ignoreError: true,
         });
         try {

test/onboard-selection.test.ts

@@ -128,9 +128,9 @@ 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);
@@ -301,8 +301,8 @@ credentials.prompt = async (message) => {
 credentials.ensureApiKey = async () => { process.env.NVIDIA_API_KEY = "nvapi-test"; };
 runner.runCapture = (command) => {
   if (command.includes("command -v ollama")) return "";
-  if (command.includes("localhost:11434/api/tags")) return "";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:11434/api/tags")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   return "";
 };
 
@@ -394,8 +394,8 @@ credentials.prompt = async (message) => {
 credentials.ensureApiKey = async () => { process.env.NVIDIA_API_KEY = "nvapi-test"; };
 runner.runCapture = (command) => {
   if (command.includes("command -v ollama")) return "";
-  if (command.includes("localhost:11434/api/tags")) return "";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:11434/api/tags")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   return "";
 };
 
@@ -536,7 +536,7 @@ const { setupNim } = require(${onboardPath});
     assert.ok(payload.lines.some((line) => line.includes("Chat Completions API available")));
   });
 
-  it("warms and validates Ollama via localhost before moving on", () => {
+  it("warms and validates Ollama via 127.0.0.1 before moving on", () => {
     const repoRoot = path.join(import.meta.dirname, "..");
     const tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), "nemoclaw-onboard-ollama-validation-"));
     const fakeBin = path.join(tmpDir, "bin");
@@ -583,9 +583,9 @@ runner.run = (command, opts = {}) => {
 };
 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";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   if (command.includes("api/generate")) return '{"response":"hello"}';
   return "";
 };
@@ -630,7 +630,7 @@ const { setupNim } = require(${onboardPath});
       payload.lines.some((line) => line.includes("Loading Ollama model: nemotron-3-nano:30b")),
     );
     assert.ok(
-      payload.commands.some((command) => command.includes("http://localhost:11434/api/generate")),
+      payload.commands.some((command) => command.includes("http://127.0.0.1:11434/api/generate")),
     );
   });
 
@@ -677,9 +677,9 @@ credentials.ensureApiKey = async () => { process.env.NVIDIA_API_KEY = "nvapi-goo
 runner.run = () => ({ status: 0 });
 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";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   if (command.includes("api/generate")) return '{"response":"hello"}';
   return "";
 };
@@ -777,9 +777,9 @@ credentials.prompt = async (message) => {
 };
 runner.runCapture = (command) => {
   if (command.includes("command -v ollama")) return "/usr/bin/ollama";
-  if (command.includes("localhost:11434/api/tags")) return JSON.stringify({ models: [] });
+  if (command.includes("127.0.0.1:11434/api/tags")) return JSON.stringify({ models: [] });
   if (command.includes("ollama list")) return "";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   if (command.includes("api/generate")) return '{"response":"hello"}';
   return "";
 };
@@ -884,9 +884,9 @@ credentials.prompt = async (message) => {
 };
 runner.runCapture = (command) => {
   if (command.includes("command -v ollama")) return "/usr/bin/ollama";
-  if (command.includes("localhost:11434/api/tags")) return JSON.stringify({ models: [] });
+  if (command.includes("127.0.0.1:11434/api/tags")) return JSON.stringify({ models: [] });
   if (command.includes("ollama list")) return "";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   if (command.includes("api/generate")) return '{"response":"hello"}';
   return "";
 };
@@ -2671,8 +2671,8 @@ credentials.prompt = async (message) => {
 credentials.ensureApiKey = async () => {};
 runner.runCapture = (command) => {
   if (command.includes("command -v ollama")) return "";
-  if (command.includes("localhost:11434")) return "";
-  if (command.includes("localhost:8000/v1/models")) return JSON.stringify({ data: [{ id: "meta-llama/Llama-3.3-70B-Instruct" }] });
+  if (command.includes("127.0.0.1:11434")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return JSON.stringify({ data: [{ id: "meta-llama/Llama-3.3-70B-Instruct" }] });
   return "";
 };
 
@@ -2780,8 +2780,8 @@ credentials.prompt = async (message) => {
 credentials.ensureApiKey = async () => {};
 runner.runCapture = (command) => {
   if (command.includes("command -v ollama")) return "";
-  if (command.includes("localhost:11434")) return "";
-  if (command.includes("localhost:8000/v1/models")) return "";
+  if (command.includes("127.0.0.1:11434")) return "";
+  if (command.includes("127.0.0.1:8000/v1/models")) return "";
   return "";
 };