agenticstore/agentic-store-mcp repository overview

⁠⚡ AgenticStore MCP: The Ultimate Open-Source AI Agent Toolkit

Supercharge your AI Assistant with 27 Powerful Model Context Protocol (MCP) Tools

Stop managing dozens of fragmented servers. Equip your AI agents (Claude Desktop, Cursor, Windsurf, LangChain, CrewAI) with a comprehensive library of powerful, free tools through a single, unified MCP server. One endpoint. Unlimited agent superpowers.

⁠🧠 What is AgenticStore MCP?

Building intelligent agentic AI systems shouldn't mean drowning in infrastructure. agentic-store-mcp is a streamlined, open-source Model Context Protocol server designed to aggregate 27 mission-critical MCP tools into one high-performance endpoint.

With AgenticStore MCP, your agents gain instant access to web searching, deep codebase analysis, security scanning, and persistent memory—all running 100% locally with zero vendor lock-in.

Most importantly, it provides an LLM Prompt Firewall, establishing a secure local proxy to intercept, scan, and sanitize any prompts leaving your system. Secure your agentic workflows with structured prompt recording and generate clear audit traces for AI usage. Flag leaked secrets, PII, and API keys, and use local LLM models (like Ollama) to sanitize your data before it reaches the cloud.

⁠⚡ Quick Start: Deploy in Seconds

Get your complete MCP tool hub running instantly. No Git login required.

⁠🐳 Option 1: Docker (Isolated Environment)

1. Pull the latest image

docker pull agenticstore/agentic-store-mcp:latest

2. Run the unified MCP server

docker run -i --rm \
  -v agentic-store-mcp_agentic-store-memory:/root/.config/agentic-store \
  agenticstore/agentic-store-mcp:latest

⁠⚡ Option 2: Run via uvx (Fastest Native Start)

You can run the complete MCP Hub Webapp and server locally using uv without cloning the repository.

1. Install uv (if you haven't already; alternatively, use pip install agentic-store-mcp):

curl -LsSf https://astral.sh/uv/install.sh | sh

2. Launch the MCP Hub Webapp directly from PyPI:

uvx agentic-store-mcp --launch-webapp

Access the rich MCP Hub UI at http://localhost:8765⁠ to visually manage connections, API keys, and auto-generate client configurations for Claude, Cursor, and Windsurf!

⁠🏆 Why Developers Choose AgenticStore MCP

• 🔒 100% Privacy-First: Everything runs locally. Your code and data never leave your machine.
• 💸 Truly Free: No accounts, no paywalls, no API subscriptions required.
• 🧠 Persistent Agent Memory: Give your AI the ultimate ability to remember facts, states, logs, and checkpoints across multiple sessions.
• ✅ The "One-Point" MCP Server: Eliminate the headache of orchestrating multiple MCP servers. Access all 27 tools through one centralized hub.
• 🛡️ LLM Prompt Firewall: Intercept, sanitize, and perform prompt recording for all data leaving your system, ensuring robust AI Data Loss Prevention (DLP).
• 🎛️ MCP Hub Webapp: A sleek, self-hosted UI to visually configure tools and manage GitHub/OpenAI API keys securely within the OS keyring.
• ⚡ Universal LLM Compatibility: Works out of the box with Claude Desktop, Cursor, Windsurf, VS Code, LangChain, AutoGen, CrewAI, and local open-source models.

⁠🛠️ The Tool Library (27 Built-In Tools)

Equip your AI client with these 4 powerhouse modules:

⁠💻 Code & Security Tools (11 Tools)

• python_lint_checker, search_code, get_file: Deep codebase mastery, fast code slicing, and navigation without external dependencies.
• repo_scanner, dependency_audit: Agent-driven DevSecOps to find leaked API keys, PII, and scan CVE vulnerabilities (OSV database).
• get_repo_info, manage_issue, create_pr: Complete GitHub integration for autonomous issue triage, code PRs, and commit statistics.

⁠🌐 Data & Search Tools (2 Tools)

• agentic_web_search: Powerful live web searches querying via a self-hosted SearXNG engine. Zero per-search costs!
• agentic_web_crawl: Extract remarkably clean markdown, word counts, and metadata directly from any URL.

⁠🧠 Memory & Agent Orchestration (12 Tools)

• memory_write, memory_read, memory_search: Robust key-value fact storage that survives system restarts.
• memory_checkpoint, restore_session: Save and load complete context snapshots and AI decision trails seamlessly.
• update_plan, update_change_log, spinup_memory: Let your agent autonomously track milestones, initialize workspaces, and append semantic release notes cleanly to your files.

⁠🔧 Meta-Tools (2 Tools)

• tool_search, configure: Let your AI dynamically discover its own available toolkit boundaries and override runtime configurations.

⁠🔗 Resources & Community

Join the revolution of open-source AI tooling.

• 🌐 Website: agenticstore.dev⁠
• ⭐ GitHub: agenticstore/agentic-store-mcp⁠ (Drop a star to support the project!)
• 📦 PyPI Package: agentic-store-mcp⁠
• 📚 Documentation: agenticstore.dev/docs⁠
• 💬 Issues & Support: GitHub Issues⁠