Launching today
Remote OpenClaw
13,000+ MCP servers, skills & plugins for AI coding agents
27 followers
13,000+ MCP servers, skills & plugins for AI coding agents
27 followers
Remote OpenClaw is the largest open directory of building blocks for AI coding agents: 13,000+ MCP servers, agent skills, plugins, and ready-made workflows for Claude Code, Codex, OpenClaw, and Hermes. Stop hunting across GitHub, npm, and scattered docs. Find, compare, and install everything in one place. Filter by client, category, and use case, copy the install command, and go. Free to browse, no login required.
This is useful for anyone building with coding agents seriously. The hard part is not just finding one MCP server, it is knowing which client it works with, how to install it, and whether there is a better option for the same use case.
Curious how you’re handling quality signals as the directory grows. Are servers ranked mostly by metadata like GitHub activity and package info, or do you plan to add usage/reliability signals too?
@aditya_harish_2002 Really appreciate this, and you nailed the hard part. Finding a server is easy. Knowing which client it works with, how to install it, and whether there's a better option for the same job is where people burn hours.
Right now ranking leans on metadata: GitHub activity, official status, package info, and client compatibility, so the well-maintained, installable ones surface first, grouped by category so you can compare like for like.
Usage and reliability signals are next: does-it-boot, install success, transport health, real adoption. Stars tell you a repo is popular, not that the server still works today. That's the gap I want to close.
Curious, what signal would you trust most, does-it-install or does-it-stay-up?
@aaron56 I’d trust does-it-install first, then does-it-stay-up once there’s enough runtime data.
If I’m trying an MCP server, the first failure mode is usually setup pain: wrong client, stale docs, missing env vars, broken package, or unclear transport. A server that installs cleanly across the clients it claims to support already clears a huge bar.
Reliability becomes the stronger ranking signal after that, but only if it is measured from real usage rather than repo popularity.
@aaron56 Makes sense. I’d probably treat install success as the entry filter, not just a ranking signal.
If it doesn’t install cleanly for the client it claims to support, it should drop hard no matter how popular the repo is. After that, uptime and real usage can decide which good options rise to the top.
Would love to see a "compatibility check" feature that warns you if an MCP server or skill hasn't been updated for your specific client version before you copy the install command, saves people from troubleshooting broken installs later.
Honestly love that you can just copy the install command straight from the listing, no extra clicks or digging through docs. Filtering by client and use case upfront is the kind of thing that sounds small but saves a ton of time when you're setting up a new agent.