feat(install): offer express install on Windows WSL

docs/get-started/quickstart.md

@@ -64,8 +64,9 @@ $ newgrp docker
 $ curl -fsSL https://www.nvidia.com/nemoclaw.sh | bash
 ```
 
-On DGX Spark and DGX Station, an interactive installer can offer express install after you accept the third-party software notice.
+On DGX Spark, DGX Station, and Windows WSL, an interactive installer offers express install after you accept the third-party software notice.
 Express install switches onboarding to non-interactive mode, allows `sudo` password prompts for required host changes, applies the suggested security policy, and selects the managed local inference path for that platform.
+On WSL, express install selects the Windows-host Ollama setup path.
 Set `NEMOCLAW_NO_EXPRESS=1` to skip the express prompt, or set `NEMOCLAW_PROVIDER` before launching the installer when you want to choose a provider yourself.
 
 The installer auto-launches `nemoclaw onboard` when it can locate the freshly-installed binary.

docs/get-started/quickstart.mdx

@@ -51,8 +51,9 @@ $ newgrp docker
 $ curl -fsSL https://www.nvidia.com/nemoclaw.sh | bash
 ```
 
-On DGX Spark and DGX Station, an interactive installer can offer express install after you accept the third-party software notice.
+On DGX Spark, DGX Station, and Windows WSL, an interactive installer offers express install after you accept the third-party software notice.
 Express install switches onboarding to non-interactive mode, allows `sudo` password prompts for required host changes, applies the suggested security policy, and selects the managed local inference path for that platform.
+On WSL, express install selects the Windows-host Ollama setup path.
 Set `NEMOCLAW_NO_EXPRESS=1` to skip the express prompt, or set `NEMOCLAW_PROVIDER` before launching the installer when you want to choose a provider yourself.
 
 The installer auto-launches `nemoclaw onboard` when it can locate the freshly-installed binary.

docs/get-started/windows-preparation.md

@@ -108,6 +108,7 @@ You can also start it yourself beforehand with `ollama serve`.
 
 You can also use Ollama for Windows.
 During onboarding, NemoClaw can use an already-running Windows-host daemon, start or restart an installed daemon, or install Ollama on the Windows host.
+If the installer offers express install on WSL, accepting it selects this Windows-host Ollama path automatically.
 When Ollama runs on the Windows host, NemoClaw detects it from WSL through `host.docker.internal` and pulls missing models through the Ollama HTTP API.
 Do not run both the Windows and WSL Ollama instances on port `11434` at the same time.
 Use one instance, or move one of them to a different port before running `nemoclaw onboard`.

docs/get-started/windows-preparation.mdx

@@ -95,6 +95,7 @@ You can also start it yourself beforehand with `ollama serve`.
 
 You can also use Ollama for Windows.
 During onboarding, NemoClaw can use an already-running Windows-host daemon, start or restart an installed daemon, or install Ollama on the Windows host.
+If the installer offers express install on WSL, accepting it selects this Windows-host Ollama path automatically.
 When Ollama runs on the Windows host, NemoClaw detects it from WSL through `host.docker.internal` and pulls missing models through the Ollama HTTP API.
 Do not run both the Windows and WSL Ollama instances on port `11434` at the same time.
 Use one instance, or move one of them to a different port before running `nemoclaw onboard`.

docs/reference/commands.md

@@ -1198,7 +1198,7 @@ These flags toggle optional behaviors during onboarding; set them before running
 |----------|--------|--------|
 | `NEMOCLAW_YES` | `1` to enable | Auto-accepts confirmation prompts (`--yes` equivalent) including in helpers like the Ollama proxy auth setup. |
 | `NEMOCLAW_NON_INTERACTIVE_SUDO_MODE` | `prompt` or empty/unset | When set to `prompt`, allows non-interactive onboarding to use prompt-capable `sudo` for host setup steps that require elevation, which can ask for a password. Empty/unset is the default and uses `sudo -n`, which fails instead of asking for a password. Any other value is rejected. |
-| `NEMOCLAW_NO_EXPRESS` | `1` to enable | Installer-only. Skips the DGX Spark and DGX Station express install prompt and continues with the normal interactive onboarding flow. |
+| `NEMOCLAW_NO_EXPRESS` | `1` to enable | Installer-only. Skips the DGX Spark, DGX Station, and Windows WSL express install prompt and continues with the normal interactive onboarding flow. |
 | `NEMOCLAW_EXPERIMENTAL` | `1` to enable | Surfaces experimental providers and flows in onboarding. |
 | `NEMOCLAW_IGNORE_RUNTIME_RESOURCES` | `1` to enable | Suppresses the under-provisioned runtime warning during preflight. Use only when you know the sandbox host meets the minimums. |
 | `NEMOCLAW_DISABLE_OVERLAY_FIX` | `1` to enable | Skips the Docker overlay-fix step during sandbox build. For environments where the fix is incompatible. |

docs/reference/commands.mdx

@@ -1165,7 +1165,7 @@ These flags toggle optional behaviors during onboarding; set them before running
 |----------|--------|--------|
 | `NEMOCLAW_YES` | `1` to enable | Auto-accepts confirmation prompts (`--yes` equivalent) including in helpers like the Ollama proxy auth setup. |
 | `NEMOCLAW_NON_INTERACTIVE_SUDO_MODE` | `prompt` or empty/unset | When set to `prompt`, allows non-interactive onboarding to use prompt-capable `sudo` for host setup steps that require elevation, which can ask for a password. Empty/unset is the default and uses `sudo -n`, which fails instead of asking for a password. Any other value is rejected. |
-| `NEMOCLAW_NO_EXPRESS` | `1` to enable | Installer-only. Skips the DGX Spark and DGX Station express install prompt and continues with the normal interactive onboarding flow. |
+| `NEMOCLAW_NO_EXPRESS` | `1` to enable | Installer-only. Skips the DGX Spark, DGX Station, and Windows WSL express install prompt and continues with the normal interactive onboarding flow. |
 | `NEMOCLAW_EXPERIMENTAL` | `1` to enable | Surfaces experimental providers and flows in onboarding. |
 | `NEMOCLAW_IGNORE_RUNTIME_RESOURCES` | `1` to enable | Suppresses the under-provisioned runtime warning during preflight. Use only when you know the sandbox host meets the minimums. |
 | `NEMOCLAW_DISABLE_OVERLAY_FIX` | `1` to enable | Skips the Docker overlay-fix step during sandbox build. For environments where the fix is incompatible. |

install.sh

@@ -117,6 +117,7 @@ bootstrap_usage() {
   printf "    NEMOCLAW_NON_INTERACTIVE=1   Same as --non-interactive\n"
   printf "    NEMOCLAW_FRESH=1             Same as --fresh\n"
   printf "    NEMOCLAW_ACCEPT_THIRD_PARTY_SOFTWARE=1 Same as --yes-i-accept-third-party-software\n"
+  printf "    NEMOCLAW_NO_EXPRESS=1        Skip express install prompt on supported platforms\n"
   printf "    NEMOCLAW_SANDBOX_NAME        Sandbox name to create/use\n"
   printf "    NEMOCLAW_ACCEPT_EXPERIMENTAL_OPENSHELL_UPGRADE=1\n"
   printf "                                 Allow automatic pre-0.0.37 OpenShell gateway upgrade\n"

scripts/install.sh

@@ -550,6 +550,7 @@ usage() {
   printf "    NEMOCLAW_NON_INTERACTIVE=1    Same as --non-interactive\n"
   printf "    NEMOCLAW_NON_INTERACTIVE_SUDO_MODE=prompt Allow sudo prompts during non-interactive onboarding\n"
   printf "    NEMOCLAW_FRESH=1              Same as --fresh\n"
+  printf "    NEMOCLAW_NO_EXPRESS=1         Skip express install prompt on supported platforms\n"
   printf "    NEMOCLAW_SANDBOX_NAME         Sandbox name to create/use\n"
   printf "    NEMOCLAW_SINGLE_SESSION=1     Abort if active sandbox sessions exist\n"
   printf "    NEMOCLAW_ACCEPT_EXPERIMENTAL_OPENSHELL_UPGRADE=1\n"
@@ -2037,7 +2038,7 @@ ensure_docker() {
   case "$(uname -s)" in
     Darwin | MINGW* | MSYS*) return 0 ;;
   esac
-  if [ -n "${WSL_DISTRO_NAME:-}" ] || [ -n "${WSL_INTEROP:-}" ]; then
+  if is_wsl_host; then
     return 0
   fi
   # Fast path: docker info works → already set up (root, or already-active group).
@@ -2117,11 +2118,31 @@ ensure_docker() {
   fi
 }
 
-# Detect DGX Spark / DGX Station from firmware (DMI first, devicetree fallback).
-# Echoes "DGX Spark", "DGX Station", or empty. Used to gate the express
-# install prompt; only platforms with a known sensible default are offered.
+is_wsl_host() {
+  if [ -n "${WSL_DISTRO_NAME:-}" ] || [ -n "${WSL_INTEROP:-}" ]; then
+    return 0
+  fi
+  if [ -r /proc/sys/kernel/osrelease ] \
+    && grep -qiE 'microsoft|wsl' /proc/sys/kernel/osrelease 2>/dev/null; then
+    return 0
+  fi
+  if [ -r /proc/version ] \
+    && grep -qiE 'microsoft|wsl' /proc/version 2>/dev/null; then
+    return 0
+  fi
+  return 1
+}
+
+# Detect DGX Spark / DGX Station from firmware (DMI first, devicetree fallback)
+# and Windows WSL from the host environment. Echoes "DGX Spark",
+# "DGX Station", "Windows WSL", or empty. Used to gate the express install
+# prompt; only platforms with a known sensible default are offered.
 detect_express_platform() {
   local model=""
+  if is_wsl_host; then
+    printf "Windows WSL"
+    return
+  fi
   if [ -r /sys/class/dmi/id/product_name ]; then
     model="$(cat /sys/class/dmi/id/product_name 2>/dev/null || true)"
   fi
@@ -2135,7 +2156,7 @@ detect_express_platform() {
   esac
 }
 
-# Prompt the user to opt into express install on Spark/Station. Sets the
+# Prompt the user to opt into express install on supported platforms. Sets the
 # non-interactive + provider/model env vars when accepted. Skipped when
 # the user already passed --non-interactive, set NEMOCLAW_PROVIDER, or has
 # no TTY.
@@ -2198,6 +2219,9 @@ maybe_offer_express_install() {
         "DGX Station")
           export NEMOCLAW_PROVIDER=install-vllm
           ;;
+        "Windows WSL")
+          export NEMOCLAW_PROVIDER=install-windows-ollama
+          ;;
       esac
       ;;
     *)

src/lib/onboard.ts

@@ -67,6 +67,9 @@ const {
 const {
   getSelectionDrift,
 }: typeof import("./onboard/selection-drift") = require("./onboard/selection-drift");
+const {
+  resolveProviderKeyFallback,
+}: typeof import("./onboard/provider-key-fallback") = require("./onboard/provider-key-fallback");
 const { isLinuxDockerDriverGatewayEnabled }: typeof import("./onboard/docker-driver-platform") = require("./onboard/docker-driver-platform");
 const {
   reconcileGatewayGpuReuseForGpuIntent,
@@ -6393,16 +6396,9 @@ async function setupNim(
         }
         selected = options.find((o) => o.key === providerKey);
         if (!selected) {
-          // Install action keys fall back to the equivalent running-provider
-          // key when the menu only emits the running entry (the install would
-          // have been a no-op anyway).
-          if (providerKey === "install-ollama") {
-            selected = options.find((o) => o.key === "ollama");
-          } else if (providerKey === "install-vllm") {
-            selected = options.find((o) => o.key === "vllm");
-          } else if (providerKey === "ollama") {
-            selected = options.find((o) => o.key === "install-ollama");
-          }
+          selected = resolveProviderKeyFallback(options, providerKey, {
+            isWindowsHostOllama: ollamaHost === OLLAMA_HOST_DOCKER_INTERNAL,
+          });
           if (!selected) {
             if (providerKey === "hermesProvider" && !hermesProviderAvailable) {
               console.error("  Hermes Provider is only available when onboarding Hermes Agent.");

src/lib/onboard/provider-key-fallback.test.ts

@@ -0,0 +1,69 @@
+// SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// SPDX-License-Identifier: Apache-2.0
+
+import assert from "node:assert/strict";
+import { describe, it } from "vitest";
+
+import { resolveProviderKeyFallback } from "../../../dist/lib/onboard/provider-key-fallback";
+
+const option = (key: string) => ({ key, label: key });
+
+describe("resolveProviderKeyFallback", () => {
+  it("maps generic install action keys to already-running provider options", () => {
+    const options = [option("ollama"), option("vllm")];
+
+    assert.equal(
+      resolveProviderKeyFallback(options, "install-ollama", { isWindowsHostOllama: false })?.key,
+      "ollama",
+    );
+    assert.equal(
+      resolveProviderKeyFallback(options, "install-vllm", { isWindowsHostOllama: false })?.key,
+      "vllm",
+    );
+  });
+
+  it("prefers Windows-host start when install-windows-ollama is requested after install already exists", () => {
+    const options = [option("ollama"), option("start-windows-ollama")];
+
+    assert.equal(
+      resolveProviderKeyFallback(options, "install-windows-ollama", {
+        isWindowsHostOllama: false,
+      })?.key,
+      "start-windows-ollama",
+    );
+  });
+
+  it("allows Windows-host install/start requests to collapse to ollama only for Windows-host Ollama", () => {
+    const options = [option("ollama")];
+
+    assert.equal(
+      resolveProviderKeyFallback(options, "install-windows-ollama", {
+        isWindowsHostOllama: true,
+      })?.key,
+      "ollama",
+    );
+    assert.equal(
+      resolveProviderKeyFallback(options, "start-windows-ollama", {
+        isWindowsHostOllama: true,
+      })?.key,
+      "ollama",
+    );
+  });
+
+  it("does not satisfy Windows-host requests with WSL or Linux local Ollama", () => {
+    const options = [option("ollama")];
+
+    assert.equal(
+      resolveProviderKeyFallback(options, "install-windows-ollama", {
+        isWindowsHostOllama: false,
+      }),
+      undefined,
+    );
+    assert.equal(
+      resolveProviderKeyFallback(options, "start-windows-ollama", {
+        isWindowsHostOllama: false,
+      }),
+      undefined,
+    );
+  });
+});

src/lib/onboard/provider-key-fallback.ts

@@ -0,0 +1,41 @@
+// SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// SPDX-License-Identifier: Apache-2.0
+
+export interface ProviderOption {
+  key: string;
+  label?: string;
+}
+
+export interface ProviderKeyFallbackContext {
+  isWindowsHostOllama: boolean;
+}
+
+export function resolveProviderKeyFallback<T extends ProviderOption>(
+  options: T[],
+  providerKey: string | null | undefined,
+  context: ProviderKeyFallbackContext,
+): T | undefined {
+  const find = (key: string) => options.find((option) => option.key === key);
+
+  switch (providerKey) {
+    case "install-ollama":
+      return find("ollama");
+    case "install-vllm":
+      return find("vllm");
+    case "install-windows-ollama":
+      // Windows-host Ollama requests may arrive from NEMOCLAW_PROVIDER before
+      // the dynamic menu knows whether Windows Ollama needs install,
+      // start/restart, or is already reachable. Collapse only to later-state
+      // entries that still point at the Windows host.
+      return (
+        find("start-windows-ollama") ||
+        (context.isWindowsHostOllama ? find("ollama") : undefined)
+      );
+    case "start-windows-ollama":
+      return context.isWindowsHostOllama ? find("ollama") : undefined;
+    case "ollama":
+      return find("install-ollama");
+    default:
+      return undefined;
+  }
+}

test/install-preflight.test.ts

@@ -484,6 +484,7 @@ exit 98
     expect(output).toMatch(/aliases: cloud -> build, nim -> nim-local/);
     expect(output).toMatch(/NEMOCLAW_POLICY_MODE/);
     expect(output).toMatch(/NEMOCLAW_NON_INTERACTIVE_SUDO_MODE=prompt/);
+    expect(output).toMatch(/NEMOCLAW_NO_EXPRESS=1/);
     expect(output).toMatch(/NEMOCLAW_SANDBOX_NAME/);
     expect(output).toMatch(/nvidia\.com\/nemoclaw\.sh/);
   });
@@ -3081,7 +3082,11 @@ exit 0`,
 });
 
 describe("installer express install prompt (sourced)", () => {
-  function runExpressPromptWithTty(answer: string, stdinMode: "pipe" | "tty") {
+  function runExpressPromptWithTty(
+    answer: string,
+    stdinMode: "pipe" | "tty",
+    platform = "DGX Spark",
+  ) {
     const tmp = fs.mkdtempSync(path.join(os.tmpdir(), "nemoclaw-express-prompt-"));
     const python =
       spawnSync("bash", ["--noprofile", "--norc", "-c", "command -v python3"], { encoding: "utf-8" }).stdout.trim() ||
@@ -3097,9 +3102,10 @@ import time
 installer = sys.argv[1]
 answer = sys.argv[2].encode()
 stdin_mode = sys.argv[3]
+platform = sys.argv[4]
 script = r'''
 source "$INSTALLER_UNDER_TEST" >/dev/null
-detect_express_platform() { printf "DGX Spark"; }
+detect_express_platform() { printf "$EXPRESS_PLATFORM"; }
 NON_INTERACTIVE=""
 NEMOCLAW_PROVIDER=""
 NEMOCLAW_NO_EXPRESS=""
@@ -3110,6 +3116,7 @@ printf "RESULT NON_INTERACTIVE=%s SUDO_MODE=%s PROVIDER=%s MODEL=%s POLICY=%s YE
 '''
 env = dict(os.environ)
 env["INSTALLER_UNDER_TEST"] = installer
+env["EXPRESS_PLATFORM"] = platform
 pid, fd = pty.fork()
 if pid == 0:
     if stdin_mode == "pipe":
@@ -3159,16 +3166,20 @@ except OSError:
 sys.stdout.buffer.write(output)
 sys.exit(exit_code)
 `;
-    return spawnSync(python, ["-c", ptyRunner, INSTALLER_PAYLOAD, answer, stdinMode], {
-      cwd: tmp,
-      encoding: "utf-8",
-      timeout: 15_000,
-      killSignal: "SIGKILL",
-      env: {
-        HOME: tmp,
-        PATH: TEST_SYSTEM_PATH,
+    return spawnSync(
+      python,
+      ["-c", ptyRunner, INSTALLER_PAYLOAD, answer, stdinMode, platform],
+      {
+        cwd: tmp,
+        encoding: "utf-8",
+        timeout: 15_000,
+        killSignal: "SIGKILL",
+        env: {
+          HOME: tmp,
+          PATH: TEST_SYSTEM_PATH,
+        },
       },
-    });
+    );
   }
 
   it("offers express install when curl-piped stdin still has a controlling TTY", () => {
@@ -3183,6 +3194,44 @@ sys.exit(exit_code)
     );
   });
 
+  it("detects Windows WSL as an express install platform", () => {
+    const result = spawnSync(
+      "bash",
+      [
+        "-c",
+        `
+source "$INSTALLER_UNDER_TEST" >/dev/null
+detect_express_platform
+`,
+      ],
+      {
+        cwd: path.join(import.meta.dirname, ".."),
+        encoding: "utf-8",
+        env: {
+          HOME: fs.mkdtempSync(path.join(os.tmpdir(), "nemoclaw-express-wsl-detect-")),
+          PATH: TEST_SYSTEM_PATH,
+          INSTALLER_UNDER_TEST: INSTALLER_PAYLOAD,
+          WSL_DISTRO_NAME: "Ubuntu",
+        },
+      },
+    );
+
+    expect(result.status, `${result.stdout}${result.stderr}`).toBe(0);
+    expect(result.stdout).toBe("Windows WSL");
+  });
+
+  it("maps Windows WSL express install to Windows-host Ollama", () => {
+    const result = runExpressPromptWithTty("\n", "pipe", "Windows WSL");
+    const output = `${result.stdout}${result.stderr}`;
+    expect(result.status, output).toBe(0);
+    expect(output).toMatch(/Detected Windows WSL/);
+    expect(output).toMatch(/Run express install/);
+    expect(output).toMatch(/Using express install for Windows WSL/);
+    expect(output).toMatch(
+      /RESULT NON_INTERACTIVE=1 SUDO_MODE=prompt PROVIDER=install-windows-ollama MODEL= POLICY=suggested YES=1/,
+    );
+  });
+
   it("skips express install without a controlling TTY", () => {
     if (process.platform === "darwin") {
       return;