Appwrite
The open-source cloud for agents and developers
4.7•31 reviews•5.5K followers
The open-source cloud for agents and developers
4.7•31 reviews•5.5K followers
Appwrite is an open-source cloud development platform designed for developers who want to get things done. Use built-in backend infrastructure and web hosting, all from a single place.
This is the 7th launch from Appwrite. View more
MCP1
Launched this week
Connect and manage all your MCP servers through a single gateway. Built by Appwrite, MCP1 lets you securely connect your AI tools, manage permissions, protect server credentials, and monitor tool usage. Works with Claude, Cursor, ChatGPT, Codex, and more. Free for individuals, with dedicated plans for teams and enterprises.
Free Options
Launch Team
Render: Get up to $100K in startup creditsCloud infrastructure for founders who want to move fast and NOT break things
Promoted
At Appwrite, we're always looking for ways to make our team more productive. As we started using AI agents across engineering, customer support, and daily operations, we ran into a problem. Connecting our agents to different tools meant managing multiple MCP servers, credentials, and permissions across different AI clients.
We wanted a simpler way to manage everything, so we built MCP1 for ourselves.
What started as an internal tool quickly became an essential part of how we work, boosting our team's productivity by 10x.
Today, we're excited to introduce MCP1 by Appwrite and make it available to everyone!
The idea behind MCP1 is simple: One URL for every MCP server.
With MCP1, you can connect all your MCP servers to AI clients like Claude Code, Cursor, ChatGPT, and Codex through a single URL. You can create personal or shared gateways, manage permissions and credentials securely, and monitor usage, all from one place.
Instead of configuring servers separately for every AI client, you can set things up once and give your agents access to the tools they need.
We built MCP1 entirely on Appwrite, and it's also part of a new series of satellite products we're building around Appwrite to help development teams become even more productive.
MCP1 is free for individuals, with dedicated plans for teams and enterprises.
You can get started at https://mcp1.com
We're proud of what we've built and excited to finally share it with the community. As always, your feedback is incredibly valuable to us, and we'll be around throughout the day to answer questions and hear what you think.
We'd love to know how you're currently using MCP servers and what you'd like to see us build next!
Thanks for your support! ❤️
@eldad_fux Does MCP1 broker server instructions for each connected MCP too? That's one gap I've seen with Cloudflare's MCP Portals that caused a bit of a different experience for clients.
@mike_ciesielka1 Not yet, and good catch. We read each server’s instructions when it syncs and show them in the dashboard, but the gateway only tells clients each server’s prefix and a one-line description. Passing them through per server is next, with tool names rewritten to match the gateway’s prefixes.
Congrats on the launch — setting things up once instead of per client is the part that sold me, that's the exact friction that's kept me from wiring up more servers. Quick question for a solo dev: is the audit log and per-person usage still useful at that scale, or is that mostly a team thing? Knowing which agent is hammering which server is what I'd want first.
@jcbuilds from my own private usage I still found it very useful when going solo. It allows me to audit my agents and actually understand what they are doing at the background, and also have piece of mind that I can have a centralize place where I can tweak permissions.
This is exactly why we add the analytics view per member aside of the usage view that gives the workspace admin eye bird view of the entire team.
That said, best way to know if it works for you is to give it a shot. Worst case, you added one more server.
with a shared gateway, can each teammate get different permissions per server, or is it one set for everyone?
@neerajnathany Yes you can set permissions per member, role, group, gateway, server or tool - full control
congrats on the launch eldad. to your question, when an agent does something weird the first thing i want is a log of which tools it called and with what input. is that visible per client in mcp1?
Picture five of us on one gateway, all hitting the same GitHub and Linear. At 2am somebody's agent blows the rate limit and everyone's stuck. Can I see usage per person, so I know whose agent it was without digging through logs?
@0x47xavier Yes, you can see usage per person. You can also set different servers for each team member using their own account. Sharing isn't the default, just one way to do things.
Per-person servers makes sense. Then my 2am question just moves. When one teammate's key hits the limit, does anyone else on the team see it, or only that person?
I use Claude Code with a few MCP servers and the tool list alone already eats a lot of context. with 10 servers behind one URL, can I choose which tools each client sees? and who refreshes an OAuth token that expires in the middle of a run?
Keeping every server credential behind one gateway instead of pasting keys into five different agent configs is what I would actually pay for, since mine are scattered across files I am afraid to touch.
@karimbenkeroum excited for you to test the platform and see if it lives up to the problem.