Mcpd

Run your agents, not your infrastructure.

Built by Mozilla AI

📚 mcpd official docs

---

mcpd is a daemon that manages your MCP servers via declarative configuration, exposing them as clean HTTP endpoints. This bridges the gap between your agents and your infrastructure, handling the messy work of lifecycle management, secret injection, and environment promotion so you don't have to.

⚙️ How it Works

Under the hood, mcpd spawns MCP servers as STDIO subprocesses and proxies requests over HTTP.

🚀 Quick Start

Prerequisites

Install the mcpd binary first, then install the runtime(s) used by the MCP servers you want to run:

• uv for servers whose package starts with uvx::
• Node.js / npx for servers whose package starts with npx::, and for mcpd inspector
• Docker for servers whose package starts with docker::, or if you plan to run mcpd in Docker

The quick start below uses the time server via uvx, so uv is required for that example.

Optional helpers used in some docs examples:

• curl to call the HTTP API
• jq to pretty-print JSON responses

If you are building mcpd from source or contributing, you will also need Go (v1.26.0+ recommended).

Installation

via Homebrew

(Works for both macOS and Linux)

Add the Mozilla.ai tap:

brew tap mozilla-ai/tap

Then install mcpd:

brew install mcpd

Or install directly from the cask:

brew install --cask mozilla-ai/tap/mcpd

Please read our docs to install via GitHub releases or local Go Binary build.

Using mcpd

# Initialize a new project and create a new .mcpd.toml file
mcpd init

# Add an MCP server to .mcpd.toml
mcpd add time

# Set the local timezone for the MCP server
mcpd config args set time -- --local-timezone=Europe/London

# Start the daemon in dev mode with debug logging
mcpd daemon --dev --log-level=DEBUG --log-path=$(pwd)/mcpd.log

Now that the daemon is running, let's call the get_current_time tool provided by the time MCP server

# Check the time
curl -s --request POST \
  --url http://localhost:8090/api/v1/servers/time/tools/get_current_time \
  --header 'Accept: application/json, application/problem+json' \
  --header 'Content-Type: application/json' \
  --data '{
  "timezone": "Europe/Warsaw"
}'

API docs will be available at http://localhost:8090/docs.

💡 Why mcpd?

Engineering teams build agents that work locally, then struggle to make them production-ready across environments. mcpd bridges this gap with declarative configuration and secure secrets management.

• Declarative & reproducible – .mcpd.toml defines your tool infrastructure
• Language-agnostic – Python, JS, Docker containers via unified HTTP API
• Dev-to-prod ready – Same config works locally and in containers

🏗️ Built for Dev & Production

| Development Workflow | Production Benefit | |-----------------------------------------------------------------------------------|------------------------------------------------------------| | mcpd daemon runs everything locally | Same daemon runs in containers | | .mcpd.toml version-controlled configs | Declarative infrastructure as code | | Local secrets in ~/.config/mcpd/ | Secure secrets injection via control plane | | mcpd config export exports version-control safe snapshot of local configuration | Sanitized secrets config and templates for CI/CD pipelines |

📦 SDKs

mcpd SDKs

| Language | Repository | Status | |------------|--------------------------------------------------------------------------|--------| | Python | mcpd-sdk-python | ✅ | | JavaScript | mcpd-sdk-javascript | ✅ |

mcpd plugin SDKs

Plugin SDKs are built using the mcpd plugin Protocol Buffers specification.

| Language | Repository | Status | |----------|----------------------------------------------------------------------------------|--------| | Go | mcpd-plugins-sdk-go | ✅ | | .NET | mcpd-plugins-sdk-dotnet | ✅ | | Python | mcpd-plugins-sdk-python | ✅ |

💻 Development

If you are developing mcpd, you will need:

• Go (v1.26.0+ recommended)

Build local code:

make build

Run tests:

make test

Validate Mozilla AI registry (when modifying registry files):

make validate-registry

Run the local documentation site (requires uv), dynamically generates command line documentation:

make docs

🧩 The Mozilla.ai Stack

mcpd is the "Tooling Layer" of the Mozilla.ai ecosystem. These tools are designed to work together or standalone.

| Layer | Tool | Function | |----------|----------------------------------------------------------------------------------|--------| | Compute | llamafile | Local LLM inference server | |Interface | any-llm| Unified Python library for LLM inference | | Logic | any-agent | Orchestration and agent loops | | Tools | mcpd | (You are here) Tool sandbox & router | | Safety | any-guardrail| Input/Output validation |

🤝 Contributing

Please see our [Contributing to mcpd](CONTRIBUTING.md) guide for more information.

📄 License

[Licensed](LICENSE) under the Apache License 2.0.

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: mozilla-ai
• Source: mozilla-ai/mcpd
• License: Apache-2.0
• Homepage: https://mozilla-ai.github.io/mcpd/

Install and usage instructions live in the source repository linked above.