Command Palette

Search for a command to run...

Home / Servers

instantaneous-windows-file-search-mcp-server

Updated 1mo ago

by clayleopardlabs

Search a million files. Instantly.

An MCP server that gives AI agents instant filesystem search on Windows (Linux and macOS in the next few days). Agents can find files by name, extension, size, or date, scope searches to a directory, exclude noise folders, and count matches across millions of files in milliseconds, without walking folders one by one.

How It Works

Windows records every file creation, rename, move, and deletion as it happens. This server maintains a live index of the NTFS Master File Table from those records, so agents can query the entire filesystem the same way a search engine queries a document index.

Typical agent use cases:

  • "Show me what changed in this project since yesterday."
  • "Find files that look like secrets or local config."
  • "List source files but ignore dependencies and build output."
  • "Count how many test files exist beside implementation files."
  • "Find old exports, duplicate downloads, or forgotten installers."
  • "Give me the project's shape before reading the code."

Search Engine (Built In)

No separate programs to install. This tool brings its own search engine, so it works out of the box:

  1. A background indexer. A small Windows service keeps a live, always-up-to-date list of everything on your drives. The first scan takes about 15 seconds on a typical PC (roughly 2.4 million files). After that it stays current on its own by watching for new, renamed, or deleted files.
  2. A Fallback Engine (instant-file-search-fallback-engine-1.5.0.1418b). Just in case. If the background indexer is ever stopped, your searches are answered automatically by the Fallback Engine instead. You never have to start or manage anything.

Searches are answered in milliseconds straight from memory. Nothing gets written to disk, and nothing leaves your machine.

Install

This tool is currently Windows-only but Linux and MacOS support should be ready in the next few days.

The recommended method

Open PowerShell:

  1. Press the Start key.
  2. Type powershell.
  3. Press the Enter key.

Run the installer:

  1. Enter the command below in the PowerShell window.

    powershell -c "irm https://raw.githubusercontent.com/clayleopardlabs/instant-file-search-MCP-server/master/scripts/install.ps1 | iex"
    
  2. Press the Enter key.

NOTE: The installer downloads the MCP server, the native indexer, and the instant-file-search-fallback-engine from the latest release on GitHub. The Rust toolchain is not required. The installer does not compile files. It also installs the OpenCode plug-in automatically. Then the sub-agents can use the same search tools.

  1. Press Yes when Windows shows the permission prompt.
  2. Restart your AI app. It can be VS Code, Cursor, Claude Desktop, or OpenCode.

NOTE: This step puts the background indexer into service. It is the only step that uses administrator rights.

Updates are staged in a new versioned folder before the installer switches the MCP client and Windows service to it. Running sessions keep their old executable until you restart the AI app; the installer never overwrites an executable that is in use. The active version and paths are recorded in %LOCALAPPDATA%\ClayLeopardLabs\instant-file-search\current.json.

NOTE: If you cannot see the permission prompt, the tools work with the fallback engine. They are slower. You can register the native indexer subsequently by running the installer in elevated mode.

NOTE: If an AI agent reports that it installed only the fallback engine, the installation is incomplete. Ask it to run the repository installer with the required UAC approval and verify with doctor.ps1 -RequireNative.

After the installation, your AI assistant has five new tools: find_files, count_files, search_status, recent_changes, and aggregate_files.

Agent installation contract

Related servers

n8n

Updated today

by n8n-io

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

204,501

mcp-server-fetch

OfficialUpdated 14d ago

by modelcontextprotocol

A Model Context Protocol server providing tools to fetch and convert web content for usage by LLMs

90,371

@modelcontextprotocol/server-everything

OfficialUpdated 14d ago

by modelcontextprotocol

MCP server that exercises all the features of the MCP protocol

90,371

mcp-server-git

OfficialUpdated 14d ago

by modelcontextprotocol

A Model Context Protocol server providing tools to read, search, and manipulate Git repositories programmatically via LLMs

90,371