Support optional CLI and MCP in Windows source archives

docs/AGENT-INTEGRATION.md

@@ -11,6 +11,30 @@ Build the CLI with `cargo build --release -p searchi-cli --locked`. Create and
 index files through the normal CLI/desktop workflow first. MCP requires an
 explicit existing database; it never creates a new empty database after a typo.
 
+Windows source archives can include the compiled CLI using
+`python tools/package_windows.py --cli target/release/searchi.exe`. In that
+archive, `searchi.exe` opens the desktop and `searchi-cli.exe` runs CLI/MCP
+commands. Keep both beside the bundled runtime helpers. The published v0.1.0
+installer does not include the CLI; this packaging option is for a subsequent
+archive and does not change that release.
+
+Packaging evidence, Windows VM 2026-10-13: 16 package tests passed in normal
+and optimized Python modes, including CLI checksum inventory and rejection of
+a mismatched CLI architecture without replacing the previous archive. The
+locked release CLI built successfully. An actual local archive containing the
+existing desktop release and this CLI passed size/SHA verification for all 40
+files. Its extracted `searchi-cli.exe mcp --help` executed successfully outside
+the build directory. This checks packaging and startup, not a new semantic MCP
+session, GPU behavior or a clean-machine install.
+
+Ignored archive: `.searchi/packages/searchi-windows-agent-preview.zip`,
+152,098,987 bytes, SHA-256
+`9721a657458f140dc9271b508c810ca52b7e4b89ccff8f1cd3edb053dbe591a0`.
+CLI SHA-256:
+`7b40a3959e5bca811f79e0a0da1e0cf77505ede703cf92c8394f79b9a6167ab5`.
+The archive is local only; the Gitea attachment size limit previously rejected
+even the smaller desktop archive. Published installer contents are unchanged.
+
 For names/content only:
 
 ```sh
@@ -105,4 +129,4 @@ cargo test -p searchi-cli --test mcp
 Default tests use genuine SQLite retrieval without downloaded model assets.
 Hybrid mode uses the same real Google service; it has no simulated embeddings.
 Native Linux/macOS MCP execution and cross-platform packaging require their own
-recorded checks; Windows CPU checks do not establish those releases.
+recorded checks; Windows CPU checks do not establish those releases.

docs/BUILDING.md

@@ -210,11 +210,23 @@ python tools/embeddinggemma2/build_runtime.py --backend auto --jobs 2
 python tools/embeddinggemma2/stage_runtime.py --backend auto
 python tools/media-runtime/provision.py --stage
 Set-Location apps/desktop
-npm run tauri -- build --no-bundle
+npm run tauri -- build --features custom-protocol --no-bundle
 Set-Location ../..
 python tools/package_windows.py
 ```
 
+To include the CLI and read-only MCP server in this source archive:
+
+```powershell
+cargo build --release -p searchi-cli --locked
+python tools/package_windows.py --cli target/release/searchi.exe
+```
+
+The additional binary is named `searchi-cli.exe`; `searchi.exe` remains the
+desktop launcher. Its model cache and database must be selected explicitly,
+as described in [agent integration](AGENT-INTEGRATION.md). This option does not
+add the CLI to the already published v0.1.0 installer.
+
 The generated `.searchi/packages/searchi-windows.zip` contains the desktop,
 matching-architecture Google worker, FFmpeg/FFprobe and notices. It excludes
 model caches, local settings and indexes. `PACKAGE.json` records SHA-256 for

tools/package_windows.py

@@ -101,7 +101,7 @@ def pe_machine(path):
         return struct.unpack_from("<H", signature, 4)[0]
 
 
-def package(root, executable, output):
+def package(root, executable, output, cli=None):
     resources = root / "apps/desktop/src-tauri/resources"
     machine = pe_machine(executable)
     architecture = {0x8664: "x86_64", 0xAA64: "aarch64"}.get(machine)
@@ -145,6 +145,10 @@ def package(root, executable, output):
     media = media_provenance(root, resources, architecture)
     expected.update(media)
     selected = [(executable, "searchi.exe"), (root / "LICENSE", "LICENSE.txt")]
+    if cli is not None:
+        if cli.is_symlink() or not cli.is_file() or not 0 < cli.stat().st_size <= MAX_MEMBER:
+            raise ValueError("CLI executable is missing, linked or oversized")
+        selected.append((cli, "searchi-cli.exe"))
     for name in ["searchi-gemma-worker.exe", "LLAMA-LICENSE.txt", "MSVC-REDISTRIBUTION.txt", "vulkan-1.dll", "VULKAN-LICENSE.txt", "media-runtime/manifest.json",
                  "media-runtime/bin/ffmpeg.exe", "media-runtime/bin/ffprobe.exe",
                  "licenses/Inter-OFL.txt", "licenses/Fluent-MIT.txt", "media-runtime/licenses/LICENSE.txt"]:
@@ -162,7 +166,7 @@ def package(root, executable, output):
     for path, name in selected:
         if not path.is_file():
             raise ValueError(f"Required package file is missing: {name}")
-        if path != executable and path != root / "LICENSE" and not path.resolve().is_relative_to(resources.resolve()):
+        if path not in (executable, root / "LICENSE", cli) and not path.resolve().is_relative_to(resources.resolve()):
             raise ValueError("Resource resolves outside the staging directory")
         if name.endswith((".exe", ".dll")) and pe_machine(path) != machine:
             raise ValueError(f"Executable architecture differs: {name}")
@@ -194,6 +198,14 @@ def package(root, executable, output):
                 "First launch downloads approximately 865 MB of verified Google model files.\n"
                 "Choose Add folder to index your own files; ordinary search stays local.\n"
                 "This is an unsigned development build, not a certified production release.\n")
+            if cli is not None:
+                archive.writestr("AGENT-SETUP.txt",
+                    "searchi-cli.exe provides command-line search and a read-only MCP server.\n"
+                    "Run searchi-cli.exe --help, or searchi-cli.exe mcp --help.\n"
+                    "Select your existing desktop index explicitly with --db; do not create another index accidentally.\n"
+                    "For semantic MCP search also supply --model-dir and --embedding-worker.\n"
+                    "Use the verified model cache prepared by the desktop and the bundled searchi-gemma-worker.exe.\n"
+                    "See https://tea.karlcom.de/karl/searchi/src/branch/main/docs/AGENT-INTEGRATION.md\n")
         os.replace(temporary_path, output)
     finally:
         temporary_path.unlink(missing_ok=True)
@@ -204,8 +216,9 @@ def main():
     parser = argparse.ArgumentParser(description=__doc__)
     parser.add_argument("--executable", type=Path, default=ROOT / "apps/desktop/src-tauri/target/release/searchi-desktop.exe")
     parser.add_argument("--output", type=Path, default=ROOT / ".searchi/packages/searchi-windows.zip")
+    parser.add_argument("--cli", type=Path, help="Include the compiled CLI/MCP client as searchi-cli.exe")
     args = parser.parse_args()
-    architecture = package(ROOT, args.executable, args.output)
+    architecture = package(ROOT, args.executable, args.output, args.cli)
     print(f"Packaged Windows {architecture}: {args.output}")
 
 

tools/test_package_windows.py

@@ -30,6 +30,24 @@ def inventory(resources):
 
 
 class PackageTests(unittest.TestCase):
+    def test_optional_cli_manifest_and_architecture(self):
+        with tempfile.TemporaryDirectory() as directory:
+            root, _, app, output = self.fixture(directory)
+            cli = root / "build/cli.exe"
+            executable(cli)
+            package(root, app, output, cli)
+            with zipfile.ZipFile(output) as archive:
+                self.assertEqual(archive.read("searchi-cli.exe"), cli.read_bytes())
+                manifest = json.loads(archive.read("PACKAGE.json"))
+                record = next(row for row in manifest["files"] if row["path"] == "searchi-cli.exe")
+                self.assertEqual(record["sha256"], hashlib.sha256(cli.read_bytes()).hexdigest())
+                self.assertIn("--db", archive.read("AGENT-SETUP.txt").decode())
+            previous = output.read_bytes()
+            executable(cli, 0xAA64)
+            with self.assertRaisesRegex(ValueError, "architecture differs"):
+                package(root, app, output, cli)
+            self.assertEqual(output.read_bytes(), previous)
+
     def fixture(self, directory):
         root = Path(directory)
         resources = root / "apps/desktop/src-tauri/resources"