Deploy MCP servers without the infrastructure headache.

Connect a repo, Docker image, or package — or install from the registry. Deploy from the dashboard, the mcpl CLI, or your AI agent. Run servers independently or combine them behind one secure Gateway. We handle runtime, security, logs, and tool analytics.

Deploy for Free Install CLI Browse registry

curl -fsSL https://mcplambda.io/mcpl/install.sh | sh

Speed

< 90 seconds

Typical path from deploy to a live MCP endpoint with a shareable URL.

Surfaces

4 ways to ship

Dashboard, mcpl CLI, registry install, and agent control plane — same platform.

Community

Build with us

Questions, demos, and agent workflows.

Join Discord → Quick start

How It Works

From sign-in to a production MCP endpoint in under 90 seconds.

Connect

Deploy

Ship from a package URL, Git repo, Docker image, or install from the registry. Dashboard, CLI, or ask your AI agent — same result.

Connect once & operate

Connect Cursor, Claude, or VS Code to one deployment or a Gateway that combines several. Stream logs, inspect tool analytics, and evolve the toolset centrally.

One MCP URL for every tool your agent needs.

Combine tools from multiple MCPLambda deployments behind one stable endpoint. Add or remove servers later without updating every MCP client.

Create a Gateway Learn about Gateways

One client configuration

Give Claude, Cursor, VS Code, or your own agent one stable MCP URL instead of configuring every deployment separately.

One access policy

Choose OAuth, API key, or public access once for the Gateway. The policy applies to every tool it exposes.

A toolset that can evolve

Add or remove running deployments later without changing the Gateway URL. Tool names are prefixed to avoid collisions.

Deploy how you actually work

Whether you live in the browser, the terminal, a catalog, or your coding agent — MCPLambda meets you there.

Dashboard

Point-and-click deploys

Connect a GitHub repo, Docker image, or package URL. Configure transport, secrets, storage, and auth — then ship from the console.

Open console

CLI

Ship from your terminal

Install mcpl, log in once, and deploy with a single command. Stream build logs, tail runtime output, and manage projects without leaving the shell.

CLI reference

Registry

Browse & install MCP servers

Search the curated registry for Official and Community servers, inspect tools and transports, then install into your project in a couple of clicks.

Preview the catalog

Agent control plane

Manage infra from your AI client

Connect Claude, Cursor, or VS Code to the MCPLambda MCP server. Deploy, search the registry, rotate keys, and check logs through natural language.

MCP server guide

From zero to a live MCP endpoint in three commands

Built for developers and agents who prefer the terminal. Install once, deploy anything from npx://, uvx://, Git, or a container image.

terminal

# install
curl -fsSL https://mcplambda.io/mcpl/install.sh | sh

# authenticate
mcpl login

# deploy a package
mcpl deploy npx://@mcp/server-time

See the product

Deployments, Gateways, registry, analytics — and a short deploy loop.

Manage every deployment from one console.

Discover servers. Install in clicks.

Browse Official and Community MCP servers, then deploy into your project — without giving up bring-your-own Git, Docker, or package deploys. 78 servers in the catalog and growing.

Install from registry Registry docs

adb-mysql-mcp-server

Official

Official MCP server for AnalyticDB for MySQL of Alibaba Cloud

stdio 3 tools

database mysql analytics

agentql-mcp

Official

Model Context Protocol server that integrates AgentQL data extraction capabilities

stdio 1 tool

web-scraping data-extraction ai

apollo-mcp-server

Official

Exposes GraphQL operations as MCP tools for AI-driven API orchestration with Apollo

streamable-http 1 tool

graphql api orchestration

arxiv-mcp-server

Community

AI assistants search and access arXiv papers through MCP with persistent paper storage

stdio 14 tools

research academic papers

astra-db-mcp

Official

Model Context Protocol server for interacting with DataStax Astra DB

stdio 16 tools

database nosql cassandra

atlassian

Community

Connect to Atlassian products like Confluence, Jira Cloud and Server/Data deployments.

stdio 98 tools

atlassian confluence jira

Prefer your own server? Deploy from Git, Docker, or a package URL with the same ops surface.

Your coding agent can deploy for you

Connect Claude, Cursor, or VS Code to the MCPLambda MCP server. With a service account token, agents can search the registry, create deployments, manage secrets, and pull logs — through natural language.

Deploy from Git, image, package, or registry mapping
Scoped service tokens — not your interactive session
Works alongside the mcpl CLI for automation

MCP server guide Create a token Agent guide →

claude_desktop_config.json example

{
  "mcpServers": {
    "mcplambda": {
      "url": "https://api.mcplambda.io/mcp",
      "headers": {
        "Authorization": "Bearer mcpl_sat_…"
      }
    }
  }
}

Create a token in Settings → API Tokens, then ask your agent to deploy.

Key Features

Deploy, discover, operate, and secure MCP servers — from first commit to production traffic.

The full stack for MCP server infrastructure.

Get Started for Free View Docs

📦

Three Deployment Flows

GitOps from a repo, pull a Docker image, or deploy a package URL (npx, uvx, pip, go). Same platform, same ops surface — pick the path that matches your workflow.

🔗

One URL for Many Deployments

Combine up to 20 running deployments behind one stable Gateway endpoint. Choose OAuth, API key, or public access, then change members without changing the URL.

📚

MCP Server Registry

Browse Official and Community servers, filter by tier and tools, then install into your project. Discovery without giving up bring-your-own-server control.

📊

Tool-Usage Analytics

See which MCP tools your traffic hits, how often they fail, and how long they take — plus a searchable invocation log per deployment. Ops clarity, not just uptime greps.

🤖

Agent-Native Control Plane

The MCPLambda MCP server lets Claude, Cursor, and VS Code deploy, inspect, and manage infrastructure through natural language — using service account tokens you control.

🛡️

Secure by Default with ToolHive

Every server runs in an isolated container with minimal permissions, network filtering, encrypted secrets, API-key or OAuth auth, and an SSE proxy that keeps container ports off the public network.

Who It's For

Purpose-built for AI engineers, platform teams, and startups who need to deploy, discover, and operate MCP servers without managing Kubernetes.

🧠

AI & Agent Developers

Turn local MCP prototypes into secure remote endpoints your agents can call. Use the dashboard, mcpl CLI, or the MCPLambda MCP server — then watch tool analytics in production.

⚙️

Tooling & Platform Teams

Expose internal APIs and services to LLMs with isolated runtimes, encrypted secrets, API-key or OAuth auth, and a registry path for vetted community servers.

🚀

AI Startups & ML Ops

Ship production-ready MCP infrastructure without a DevOps hire. Projects, compute units, live logs, and team seats scale with you from free tier to enterprise.

Ready to ship your first MCP server?

Free tier included. Deploy from the console, mcpl CLI, or your AI agent.

Get Started for Free View Docs

Frequently Asked Questions

Everything you need to know about using MCPLambda

What is MCPLambda?

MCPLambda is a Platform-as-a-Service for AI agent infrastructure. Deploy and manage MCP servers from a Git repo, Docker image, package URL, or the server registry — without running Kubernetes yourself.
Do I need DevOps experience?

No. MCPLambda abstracts orchestration, networking, storage, and security. You configure the server; we run the runtime. Power users can still use the mcpl CLI and service tokens for automation.
Which deployment methods are supported?

Four paths into production: GitOps from a GitHub repository, direct Docker/OCI images, package URLs (npx, uvx, pip, go), and one-click install from the MCP server registry.
Is there a CLI?

Yes. Install mcpl with a one-liner, run mcpl login, then mcpl deploy to ship servers, stream logs, manage secrets, and search the registry from your terminal — ideal for developers and AI agents.
What is the MCP server registry?

A curated catalog of Official and Community MCP servers you can browse, filter, and install into your project. You still keep full bring-your-own-server control via Git, Docker, or packages.
Can I combine multiple MCP servers behind one URL?

Yes. An MCPLambda Gateway combines tools from up to 20 running deployments behind one stable MCP endpoint. Choose OAuth, API key, or public access for the Gateway, then add or remove member deployments without changing its URL.
How do monitoring and analytics work?

Every deployment gets live runtime logs plus tool-usage analytics: call volume, error rates, tool latency (p50/p95), and a searchable invocation history so you can see which tools agents actually hit.
Can my AI agent manage MCPLambda for me?

Yes. Connect Claude, Cursor, or VS Code to the MCPLambda MCP server with a service account token. Agents can deploy servers, search the registry, manage secrets, and check logs through natural language.
How secure is MCPLambda?

Every MCP server runs on ToolHive in an isolated container with minimal permissions, network filtering, and encrypted secrets. Choose API-key, OAuth, or no auth per deployment. An SSE proxy keeps container ports off the public network.
How long does a typical deployment take?

Most deployments complete in under 90 seconds. From sign-in (or mcpl login) to a running endpoint with a shareable deployment URL.

Resources & Guides

Fundamentals, production deploy, use cases, and platform comparisons — from first tool to remote MCP at team scale.

Fundamentals

MCP Client vs Server vs Host: The Mental Model That Unlocks Everything

Host, client, and server roles in the Model Context Protocol — as defined in the official MCP architecture docs — with plain-English examples.

Read guide 10 min read

Fundamentals

Build Your First MCP Server (Official Quickstart Path)

How to build a first MCP server using official Model Context Protocol docs and SDKs — tools, stdio logging rules, Claude Desktop config, and links to complete weather-server tutorials.

Read guide 10 min read

Deployment

Production-Ready MCP with mcplambda.yaml (Git Deploy Config)

Build a production-ready MCP server using mcplambda.yaml — what the file is, how MCPLambda processes it, and a full Git deploy walkthrough.

Read guide 10 min read

Use cases

MCP for Coding Agents: Repo, CI, Issues, and Deploy Tools

Use MCP with Cursor, Claude, and VS Code for engineering workflows — official transport and client guidance, security, progressive tool discovery, and remote team sharing.

Read guide 10 min read

Comparison

MCPLambda vs Vercel: Real MCP hosting vs bolting MCP onto Functions

Why serious MCP builders choose MCPLambda over Vercel Functions for production servers. Compare mcp-handler friction, tool analytics, Docker limits, and when Vercel still makes sense.

View comparison 6 min read

Comparison

MCPLambda vs AWS, GCP & Azure: Ship MCP in minutes, not a cloud project

Why serious MCP builders choose MCPLambda over raw AWS, GCP, or Azure hosting. Compare setup friction, IAM tax, tool analytics, and ops burden — and when a hyperscaler still makes sense.

View comparison 7 min read

Browse all guides All comparisons