"Local small language model for filesystem and network tasks" matching MCP servers:

Evaluates task suitability for local models before cloud API calls, routing to Ollama or similar to reduce costs.

Provides a local MCP server for AI-powered web crawling on Windows, enabling agents to directly crawl and extract web content without requiring WSL.

MCP Server that enables LLMs to interact with the local filesystem.

Local-first MCP server that lets AI agents design Cisco-style networks and generate L1/L2/L3 topology diagrams, device tables, and AI-ready context files.

A Model Context Protocol server providing network diagnostic tools for AI agents, including connectivity testing, batch operations, local network info, and pcap analysis.

An MCP server that seats multiple LLMs as a council, letting your assistant query them in parallel or in sequence, relay answers for cross-critique, and merge conclusions within one conversation.

Enables AI assistants to understand lab topologies, follow operational principles, and interact with real network devices through observable terminals while keeping private access information hidden.

MCP server that pings 130+ free coding LLM models across 17 providers in real-time, ranks them by latency, and helps AI agents pick the fastest available model.

Provides read-only access to UniFi Network consoles for sites, devices, clients, vouchers, networks, Wi-Fi, and firewall configuration, with optional opt-in tools for controlled write operations and legacy controller support.

Local MCP server that lets Claude Code delegate bounded, read-only tasks to external AI models (Codex, Grok, local OpenAI-compatible servers) via CLIProxyAPI.

Enables AI coding agents to read Microsoft Planner and To Do tasks across multiple M365 tenants with proper authentication, and supports optional write operations limited to tasks the agent created.

Enables reading, writing, appending, and listing files, restricted to explicitly allowed directories with safety checks against path traversal and symlink escapes.

Provides MCP tools for governed multi-vendor network device operations, including configuration management (backup, diff, merge, replace, rollback) and read-only queries (facts, interfaces, BGP, LLDP, ARP) via NAPALM, with optional NetBox source-of-truth integration.

Exposes Linux filesystem operations as tools for LLM agents, with configurable root directory and read-only mode, built with FastMCP.

Persistent 3D/2D spatial world model MCP server for AI agents. Entity tracking, object permanence, AABB collision simulation & view frustum projection via local SQLite.

Provides local hybrid search over files by combining semantic vector and full-text BM25 retrieval, with MCP stdio/HTTP support, OAuth 2.1 authentication, and tools for reading and converting documents to Markdown.

An MCP server that turns the Hugging Face Hub into a conversational model picker with a tool-use eval harness to prove the agent actually uses it correctly.

Enables Claude to fetch web pages from the user's own local network connection instead of AI datacenter IPs, including optional authenticated sessions, with verdicts for bot challenges, login walls, and other failures.

Enables AI clients to query, aggregate, and visually inspect a live Autodesk Revit model, reading it without model-changing transactions by default. Model mutations are opt-in and require two explicit gates, with dry-run previews and verification.

Delegates mechanical text-to-text tasks to a local LLM to save Claude subscription quota, reading files server-side so large content never enters Claude's context.