shutup-mcp
UnclaimedZero-config MCP proxy that intelligently filters tools for AI agents, drastically reducing token usage.
Visit WebsiteFreemiumVisit Website
TL;DR - shutup-mcp
- Reduces AI agent tool visibility by 99% for focused task execution.
- Achieves 98% token reduction and 85% faster response times for AI agents.
- Operates locally with zero configuration and supports offline embedding models for privacy.
Pricing: Free plan available
Best for: Growing teams
What is shutup-mcp?
shutup-mcp is a zero-configuration proxy designed to sit between your AI agent (specifically Anthropic's Claude Desktop) and multiple Model Context Protocol (MCP) servers. Its primary function is to address the problem of AI agents being overwhelmed by a vast number of available tools, which leads to increased token usage, slower response times, and reduced tool selection accuracy.
The proxy works by reading your Claude Desktop configuration, aggregating tool lists from all connected MCP servers, and then using a local embedding model (either sentence-transformers or Ollama) to build a searchable index. When the agent requests tools, shutup-mcp intercepts the call and returns only the most relevant 3-5 tools based on the user's current intent, effectively hiding the vast majority of irrelevant options. This results in significant token reduction, faster responses, and improved tool selection accuracy, making AI agents more focused and efficient without requiring any API keys or complex setup.
Available on: Web
Pros & Cons
Pros
- Significantly reduces token usage for AI agents.
- Improves AI agent response times.
- Enhances tool selection accuracy by focusing the agent.
- Operates completely locally, ensuring privacy with no API keys.
- Requires zero configuration with existing Claude Desktop setups.
Cons
- Currently specific to Claude Desktop MCP servers.
- Requires local installation and command-line execution.
- Limited to Python environment and dependencies.
Key Features
Multi-Server Aggregation from claude_desktop_config.jsonDynamic Reload of tool index upon config changesLocal & Private Embedding with sentence-transformers (default) or OllamaIntent-based Tool Filtering (returns top-K relevant tools)Zero-configuration setup with existing Claude Desktop MCPCommand-line interface for execution and configuration
Pricing Plans
Free TrialPricing checked Jul 27, 2026
Free
$0 USD per month
- Unlimited public/private repositories
- Dependabot security and version updates
- 2,000 CI/CD minutes/month (Free for public repositories)
- 500MB of Packages storage (Free for public repositories)
- Issues & Projects
- Community support
- GitHub Copilot Access
- GitHub Codespaces Access
Team
$4 USD per user/month
- Everything included in Free
- Access to GitHub Codespaces
- Repository rules
- Multiple reviewers in pull requests
- Draft pull requests
- Code owners
- Required reviewers
- Pages and Wikis
- Environment deployment branches and secrets
- 3,000 CI/CD minutes/month (Free for public repositories)
- 2GB of Packages storage (Free for public repositories)
- Web-based support
- GitHub Secret Protection
- GitHub Code Security
Enterprise
Starting at $21 USD per user/month
- Everything included in Team
- Data residency
- Enterprise Managed Users
- User provisioning through SCIM
- Enterprise Account to centrally manage multiple organizations
- Environment protection rules
- Repository rules
- Audit Log API
- SOC1, SOC2, type 2 reports annually
- FedRAMP Tailored Authority to Operate
Reviews
Be the first to review shutup-mcp
Your take helps the next buyer. Verified LinkedIn reviewers get a badge.
Write a reviewBest shutup-mcp Alternatives
Top alternatives based on features, pricing, and user needs.
OpenAI PlatformPaid
API access to powerful AI models
Hugging FaceFreemium
Open-source AI models, datasets, and tools for collaborative ML
LangChainFreemium
Build LLM-powered applications
LlamaIndexFreemium
Data framework for LLM applications
HaystackFreemium
The open-source AI framework for building and orchestrating production-ready RAG and agentic AI applications.
DSPyFree
Program large language models with structured code, not brittle strings.
Explore More
shutup-mcp FAQ
How does shutup-mcp achieve a 98% token reduction for AI agents?
shutup-mcp achieves this by acting as a proxy that filters the vast number of tools available to an AI agent. Instead of presenting all 80,000+ tools, it uses a local embedding model to understand the user's intent and only returns the 3-5 most relevant tools to the agent, drastically cutting down the context window size and thus token usage.
Can shutup-mcp be used with AI agents other than Claude Desktop?
Based on the provided information, shutup-mcp is specifically designed to parse
claude_desktop_config.json and interact with Model Context Protocol (MCP) servers used by Claude Desktop. Its current functionality is tightly coupled with this ecosystem, suggesting it may not directly support other AI agents without modifications.What are the privacy implications of using shutup-mcp?
shutup-mcp is designed with privacy in mind. It operates locally and offers two embedding backend options:
sentence-transformers (which downloads a local model) and ollama (which uses a locally running Ollama instance). Both options ensure that your data and tool indexing remain entirely on your local machine, with no external API calls or data sharing.How does shutup-mcp handle changes to my MCP server configuration?
shutup-mcp includes a dynamic reload feature. It continuously watches your
claude_desktop_config.json file for changes. If you add or remove MCP servers from your configuration, shutup-mcp automatically detects these changes and rebuilds its internal tool index without requiring a restart.Source: github.com