A local proxy that wraps your MCP servers and checks each tool call against your policy and the current state before any server tool is called. When it blocks, the agent gets back a reason it can act on.
Website: https://mcpboundary.com · Docs: https://mcpboundary.com/docs/mcp-boundary
Wrap an existing local MCP server — your agent keeps the same tools, every call passes the boundary first.
Choose which tools the agent sees — expose, hide, or keep dashboard-only.
Allow, block, or hold for review — with a structured reason the agent can act on.
Restrict specific arguments, not whole tools — e.g. allow only certain recipients or IDs.
Limit response sizes and timeouts — and throttle repeated calls to contain runaway loops.
Bind writes to observed state — don't run a write if the world changed since the read.
Inspect every call, decision, reason, and outcome in a local dashboard.
Related MCP server: protect-mcp
MCP Boundary sits between your MCP client and your local MCP servers. Your agent connects to MCP Boundary as a server entry; MCP Boundary checks each tool call against your policy and the current state, then forwards only admitted calls to the real server. When it blocks, the agent gets back a reason it can act on — re-check state, narrow scope, or stop retrying.
Windows and Linux builds with SHA-256 checksums (current release v0.2.3): → https://mcpboundary.com
Quick start from an extracted package — Windows: .\mcpboundary.exe quickstart email · Linux: ./mcpboundary quickstart email
Getting started, policy examples, tested servers — https://mcpboundary.com/docs/mcp-boundary
How it works — https://mcpboundary.com/how-it-works
MCP Tool Rules (writing policies) — https://mcpboundary.com/mcp-tool-rules
MCP Boundary works best with local command-based stdio MCP servers. Only calls routed through MCP Boundary are covered. It is not an enterprise security gateway, a DLP system, a prompt-injection detector, or a replacement for code review, database permissions, or email approval processes.
MCP Boundary is free to download and use, distributed under its own license and terms — see LICENSE.md and TERMS.md in the release package. All rights reserved.
Built by Impact Boundary Labs.
This server cannot be installed
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Security firewall for AI agents — scans MCP calls for injection, secrets, and risks.
See, price, and control every tool call your AI agents make: policy checks, cost, and audit tools.
Runtime permission, approval, and audit layer for AI agent tool execution.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/impact-boundary-labs/MCPBoundary'
If you have feedback or need assistance with the MCP directory API, please join our Discord server