"How to work with local files" matching MCP servers:

Enables ChatGPT to read and edit local text files, apply precise string replacements, and run Bash commands on a user's own macOS or Linux machine through a secure local tunnel without exposing public ports.

A local AI agent workspace that allows users to interact with simulated data files via custom slash commands and context mentions, powered by Google GenAI SDK and MCP.

Enables ChatGPT to work directly on local projects by reading files, editing code, running tests and commands, inspecting Git state, and managing durable workflows, with optional delegation to local Codex agents for complex tasks.

Enables ChatGPT Desktop to read, search, edit and move files on your own computer, produce Word, Excel, PowerPoint, PDF and chart outputs, launch and manage terminal programs and processes, and run work on other machines you own by name over SSH. Everything runs locally under your own setup, with a Docker-isolated, offline document mode or a full-permission mode for file and code operations.

Enables supervised PowerShell execution on Windows, allowing the DeepSeek agent to run local commands and manage files safely through a single tool.

Enables AI agents to work with multiple local projects on Windows, providing file system access and PowerShell command execution, with project auto-detection and support for various dev tools.

Enables cloud agents to securely operate local machine resources (files, commands, screenshots) via standard MCP protocol.

Enables Claude Desktop and Claude Code to list, search, and read files, run one terminal command, query SQLite, and manage named scheduled tasks without calling the Claude API.

Enables a Notion MCP Agent to work as a local coding agent by exposing local files, shell, git, and patch-style editing over MCP, plus delegated long-running tasks via local Codex or Claude. It also supports managing local MCP services and Cloudflare tunnels through a native macOS console.

Enables ChatGPT to operate a local workstation through a secure outbound tunnel, exposing 44 MCP tools for files, git, projects, processes, browsers, native apps, Office documents, and optional Codex delegation across macOS, Linux, and Windows.

An unofficial self-hosted MCP server that enables ChatGPT to run commands on your computer through a secure tunnel, manage processes, and work within a specified project directory.

Turns your local machine into an MCP server with a GUI, allowing agents to execute file operations, run commands (including admin), manage background processes, and query databases, with optional ngrok exposure.

Enables ChatGPT Web Developer Mode to securely read, search, list, and edit files inside user-approved local folders, and optionally run approved developer commands on the user's Windows machine.

Enables cloud services with MCP support, such as Notion AI, to read, search, edit, create, and delete files within a single designated folder on your computer, and to run only allowed commands.

Enables Claude to interact with local workspace files through MCP, including listing files, summarizing CSV datasets, generating plots, executing Python scripts, and running allowlisted shell commands from a chat interface.

Enables exploration and search of local filesystems using glob pattern matching to find files and grep to search for text patterns within files.

Enables orchestrating models to delegate reading and answering questions about local files to a local LLM, avoiding sending file contents through the orchestrator's context.

A lightweight MCP server that enables AI assistants to execute local development tools and retrieve system status with low latency over stdio or HTTP.

Enables AI agents to run WP-CLI commands on WordPress sites, including byte-faithful post and meta copying with checksum verification, over local Docker, Pantheon Terminus, or SSH.

A production-ready MCP server that enables AI assistants to execute shell commands, manage files, monitor system resources, and automate complex workflows with advanced features like stock tracking and web automation.

Enables MCP clients to run Hayabusa detection scans over Windows event log (.evtx) files for forensic analysis and threat hunting.

Provides blocking await tools to poll commands, URLs, or files until a condition is met, eliminating the need for sleep loops in agent workflows.

Enables an AI assistant to read and control your real, logged-in Safari browser tabs, with a local model gating every write action for safety.

A self-hosted MCP server providing a governed interface for AI agents to interact with local Docker infrastructure, featuring a two-phase confirm protocol for state-changing operations and append-only audit logging.

Enables cloud-based MCP clients to delegate privacy-sensitive or large-output tasks to a local LLM, returning sanitized placeholders or digested summaries so private data never leaves the user's network.

Provides agents with real Linux VMs to run commands, manage files, expose ports, and track costs, all through an ephemeral sandbox service.

Enables local forensic analysis of files by orchestrating system binaries (file, exiftool, strings, Volatility) via safe subprocess execution.

Enables an AI agent to run shell commands and move files inside a WSL distro on the same Windows machine via wsl.exe, with no sshd, key, or port required. Each pre-registered distro-and-user entry enforces guard policies limiting dangerous commands, Windows interop, write targets, and accessible paths.

Developers: Search your zsh, bash, or fish shell history from Claude Code, Cline, Cursor, Zed, or any MCP client using tools like search_history (full-text with timestamp/CWD/exit code), recent_in_dir, failed_commands, and command_chains for multi-step sequences. Reindex after new activity. Local-only SQLite FTS5 with secrets redacted before storage.

Enables handing one self-contained job to the Claude Code CLI on the local machine and getting back its final message plus a manifest of the artifacts it produced, without letting Claude write into the user's project. Reads are limited to named directories and all writes stay in a throwaway run directory, so the user reviews the listed files and copies in only what they approve.