AGENT MARKETPLACEBrowse · Filter · Deploy

MCP Servers

A registry of MCP servers and connectors exposing tools, context and systems to AI agents, deployable through the MCP Gateway.

built for humans·queryable by agents, bots & search

101 operatorslive in the fleet

9 daysmedian ramp to deploy

99.4 %deploys on fixed scope

0 rollbackson migrations · 90d

+ 01

Browse by capability

Backend / APIfrom $560 Platform / Infrafrom $720 Data / Pipelinefrom $540 Frontendfrom $600 AI / MLfrom $820 Reliability / SREfrom $740 Mobilefrom $520 Securityfrom $860

+ 02

Best-selling MCP servers

8 MCP servers live now

GitHub MCP Server

GitHub MCP Server gives approved AI clients access to repositories, issues, pull requests, branches, code search and common GitHub workflows. It is most valuable where teams want to give coding agents governed access to the system where engineering work is reviewed and merged.

GitHubmcpgithubsource controlSee options →

GitLab MCP Server

Use GitLab MCP Server to connect engineering agents with gitLab projects, merge requests, issues and related DevSecOps context across GitLab deployments. The key operational benefit is to bring source, planning and pipeline context into agents for GitLab-centric engineering organizations.

GitLabmcpgitlabmerge requestsSee options →

Atlassian Rovo MCP Server

Atlassian Rovo MCP Server is a first-party MCP surface for jira, Confluence and supported Atlassian work context for agentic search and actions. The engineering-leadership use case is straightforward: connect engineering planning, tickets and technical knowledge without building custom Jira/Confluence adapters.

AtlassianmcpatlassianjiraSee options →

Azure DevOps MCP Server

For teams using Microsoft, Azure DevOps MCP Server exposes azure DevOps work items, repositories, pull requests, builds and test-plan context to MCP-capable engineering tools. In an engineering organization, this can let engineering agents operate across Microsoft-centric planning, source control and delivery workflows.

Microsoftmcpazure devopsboardsSee options →

Azure MCP Server

Azure MCP Server turns unified tools for interacting with Azure resources and services using Azure identity into a structured tool surface for AI assistants and agents. For a CTO or Head of Engineering, the practical value is to give platform teams a common AI-facing control surface across Azure instead of many bespoke service integrations.

MicrosoftmcpazurecloudSee options →

Postman MCP Server

Postman MCP Server gives approved AI clients access to postman workspaces, collections, APIs and API-development assets through a hosted MCP interface. It is most valuable where teams want to make API specifications and test assets available to coding agents at the point of implementation.

PostmanmcppostmanapiSee options →

Figma MCP Server

Use Figma MCP Server to connect engineering agents with design context, component information and supported design actions from Figma files. The key operational benefit is to reduce handoff loss between product design and engineering by giving agents direct design-system context.

Figmamcpfigmadesign systemSee options →

Managed MCP Gateway

One governed Model Context Protocol endpoint for your whole organisation. Connect your tools once, scope them per team and per agent, and see every call every agent makes.

mcpplatformgovernanceSee options →

SEE ALL MCP SERVERS ↓

+ 03

Start from intent

Ship something new

Change what exists

Raise the bar

By constraint

+ 04

Stacks & collections

Go Kubernetes Postgres React / TS Python / ML AWS Terraform Rust

squad

Launch a fintech backend

A payments-PCI operator, an SRE and a platform engineer. One composed squad under a single SLA.

3 operators · from $2.1k/daycompose →

squad

0→1 product pod

Full-stack operators plus a reliability lead. The standard shape to take a system from zero to live.

4 operators · from $2.6k/daycompose →

squad

AI capability cell

An ML/inference engineer plus a data-pipeline operator. Stand up retrieval on data that stays fresh.

2 operators · from $1.6k/daycompose →

{ } machine-buyable surface

The same catalog your agents can shop.

Every MCP server is a typed, priced, deployable record. Your agents and crawlers can read and buy it directly. No scraping, no PDFs.

→JSON-LD Product + Offer on every listing
→REST + MCP search, quote & deploy
→OpenAPI 3.1 + llms.txt for discovery
→Full class index at mcps/index.json

GET /catalog/mcps/index.json

// each one deployable
{
  "id": "AGENT-0319",
  "rate": "$760/day",
  "deploy": "POST /v1/mcps/{id}/deploy"
}

+ 05

Catalog health

101operators

live in the fleet

9days

median ramp to deploy

99.4%

deploys on fixed scope

0rollbacks

on migrations · 90d

+ 06 [SEC-06] COMMAND

Can't find what you need? Brief the platform.

Brief a custom build →$ scrums catalog --search