"Marking Linux terminal commands as safe or unsafe for local execution" matching MCP servers:

Enables AI agents to query OpenRouter model information including prices, ELO rankings, context, and perform comparisons.

MCP server that provides local, de-identified access to a demo EMR database using a private LLM for de-identification, ensuring raw patient data stays local.

MCP server that provides OpenRouter model pricing data, enabling price lookups, trending/cheapest lists, and model searches without an API key.

Self-hosted Linux administration over MCP: 38 tools for processes, disks, network, systemd, journal and files, served over HTTPS. Each call runs as the caller's own OS user; root is granted per tool, limited to paths, networks and sysctl keys.

BTC Decision Terminal for AI Agents — live vault-backed Hyperliquid signals, on-chain proof, cross-chain swap. MCP server hosted at https://hypernatt.com/mcp/protocol. 11 tools. Verify in real time.

Enables AI agents to read local business data and request database mutations, while requiring human approval before updates or deletions are executed. It provides read-only tools, approval workflows, and audit logging to prevent autonomous destructive changes.

Enables coding agents to add a bounded semantic-judgment layer for routing, ranking, extraction, verification, and escalation, returning typed signals and review recommendations.

Local-first MCP server for controlling a loopback ComfyUI instance from same-machine or authenticated LAN clients.

Enables local llama.cpp inference and dotbot agent workflows through MCP, providing one-shot chat completions and agentic tool loops.

Routes code generation tasks between local models and Claude Cloud, optimizing cost by handling simple tasks locally and reserving cloud thinking for complex tasks.

Enables running MCP tools against local MLX models on your Mac, with hardware-aware configuration, CLI streaming, and a dashboard for routing and monitoring.

Enables MCP clients like Claude Code to delegate coding tasks to a local model through a configurable coding-agent CLI backed by any OpenAI-compatible endpoint, with tools for health checks, task execution, and context-aware file modifications.

Lets MCP clients delegate private or bulk text tasks to a local OpenAI-compatible LLM while keeping data on the machine.

Enables Codex to consult Claude Code for a second opinion through a local stdio MCP server, exposing the ask_claude tool for text-only reviews and recommendations using existing Claude Code authentication. Codex can request findings and tradeoff analysis without allowing Claude to edit files or run commands.

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

Two self-hosted MCP servers: manage a local model machine (Ollama pull/switch, LoRA training) and bridge to local Ollama/vLLM for pure language processing tasks (writing, summarizing, classifying, extraction) without giving the calling agent tools or file access.

Lets a text-only LLM read images and PDFs by delegating to a vision model running on your own GPU.

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.

A thin MCP server that delegates lightweight tasks from Claude Code or any MCP-compatible client to local or cloud LLMs via LiteLLM, supporting models like Ollama and cloud APIs as subagents.

Unified MCP server for managing local model runtimes (Ollama, LM Studio, etc.), enabling provider-agnostic discovery, lifecycle management, hardware-fit checks, and delegated inference.