About AIAgentAlts

AI agents get useful when they can reach your tools: a database, a browser, an issue tracker, a payment system. The Model Context Protocol (MCP) is the common way to connect them, and skills (folders with a SKILL.md file) teach agents how to do a particular job well.

There are now tens of thousands of MCP servers in public registries, and most of them are experiments. AIAgentAlts lists a smaller set: servers published by the company behind the service, the MCP project's reference servers, and community projects that are widely used and still maintained.

Every listing shows the facts you need to choose: whether it's official or community-built, its licence, how many GitHub stars it has, when it was last updated and how to install it.

Being listed here isn't a security review. MCP servers and skills run with whatever access you give them, so check the code, the permissions and the publisher before connecting one to anything important.

How listings are chosen and checked

Listings come from the official MCP Registry (registry.modelcontextprotocol.io), the modelcontextprotocol/servers repository, vendors' own documentation and well-known public skills repositories on GitHub. Descriptions are written by us, not copied from the projects' marketing.

“Official” means the server is published by the company that runs the service it connects to (for example GitHub's GitHub server), or by the MCP project or Anthropic. Everything else is “Community”. Install hints come from the MCP Registry entry or the project's README.

Stars, last-update dates and licences are read from GitHub on a schedule. Where a listing points at a folder inside a larger repository, stars are for the whole repository and the date is for that folder. Every URL is checked, and listings whose links stop working are removed.

Connecting an agent to real systems carries risk. Many servers can read or change data, so give them the narrowest access that works, prefer read-only modes, and read the project's security notes. Top Pick listings are clearly labelled and never change the facts shown.

Spotted something wrong or missing? Tell us.