Disco.dev is the easiest way to spin up, test, and remix MCP servers. Zero setup, no infra. Open-source and community-driven, Disco makes MCP fun, fast, customisable and accessible in one central hub.
At StackOne, every team has been building LLM-based automations, but we kept hitting the same wall: connecting agents to our vendor stack was the hardest part of making them actually useful.
⚡ MCP is emerging as the standard for tool calling, but today: - Hosting your own servers = infra headaches - No central place to discover, or share and collaborate on new servers - you can’t easily customize or edit existing servers & tools
**We built Disco.dev because making AI agents actually useful is still way harder than it should be.**
Disco is your personal MCP hub: 🪩 Zero-setup servers → spin one up in a click, no infra 👉 Open source connectors → [github.com/StackOneHQ/mcp-connectors](http://github.com/StackOneHQ/mcp...) 💻 Local + CLI support → run servers yourself if you prefer 🎛 Remixable servers → browse, fork, and customize with ease
✨ What’s Next Playground → test servers and tools directly in the app Vibe-code tools → mix & match servers, tweak behaviour, create personal tools (think Lovable for MCPs)
What’s unique? We’re the first to make remote MCP servers truly plug-and-play so you can drop them into any client with zero friction. And because they’re open to the community, Disco becomes the one place to find, build, and share the best servers, with enough guardrails and good DX to keep quality high.
We’d love to hear: What’s the first tool you’d want your favourite agent/LLM client to use if spinning it up was dead simple?
💜 Guillaume & the StackOne team
PS: For anyone based in London, we'll be doing a Disco MCP hack evening with £5k Prize money: https://lu.ma/duwuc5hn
@anwarlaksir Good question, once you create a Disco server, you do need to add it wherever you'd like to use it.
We'll look into having a way for all servers created to be auto-discovered but given the quantity of tools/MCP servers that most clients can deal with it's probably best to selectively link the servers you created to the various LLM clients that need them (eg. you might setup a Disco Jira MCP server with 4 tools to use within claude, while you won't care to connect it to let's say Cursor and you might want to connect another one, eg. Parallels instead to Cursor)
Report
@s1guillaume MCP tool calling is powerful but the setup complexity has been holding back adoption. Making remote MCP servers truly plug-and-play is brilliant! Congrats on the launch!
Amazing work StackOne team! I really like the smooth DX that makes adding the MCP server into one of the supported MCP clients so seamless. Very curious to try it out.
@ed_forson Thanks!!!!! The plan is to make the integrations super easy for AI to generate. Then expose orgs and teams so you can share your servers internally. It's what we have at StackOne and itss so good.
First and foremost we're gonna add more integrations & tools but also feature so anyone can add tools without having to work on the open sourced repo but instead via a chat-based/lovable like experience, from prompt to new integration/tool available in your Disco account in < 30 seconds.
We'll also showcase Disco at the next AI Demo Day event we're doing (in Stockholm this time around) on the 3rd of September!
generate connectors via in app chat, from prompt to available in < 30 seconds.
private servers per user and sharing per organization (how we use it internally at StackOne)
hack next week with £5k prize money and we'll also demo Disco at the next AI Demo Day in Stockholm on the 3rd of September!
Report
As someone who's been wrestling with MCP servers for months, this is a breath of fresh air. The "zero infrastructure" promise actually delivers. You can spin up connections to Slack, Linear, and GitHub repos in under 5 minutes. No Docker containers, no environment configs, no deployment headaches.
I really like the "remix-ability" aspect. Being able to browse existing servers, fork them, and customize the tools for a specific use case is brilliant. It's like having a community-powered library of AI integrations that actually work out of the box. The fact that it's all open source (just starred the repo!) makes it even better because I can see exactly what's happening under the hood (and contribute back).
Quick question for the team: What about support for custom authentication, for example, for enterprise tools?
Disco.dev
At StackOne, every team has been building LLM-based automations, but we kept hitting the same wall: connecting agents to our vendor stack was the hardest part of making them actually useful.
⚡ MCP is emerging as the standard for tool calling, but today:
- Hosting your own servers = infra headaches
- No central place to discover, or share and collaborate on new servers
- you can’t easily customize or edit existing servers & tools
**We built Disco.dev because making AI agents actually useful is still way harder than it should be.**
Disco is your personal MCP hub:
🪩 Zero-setup servers → spin one up in a click, no infra
👉 Open source connectors → [github.com/StackOneHQ/mcp-connectors](http://github.com/StackOneHQ/mcp...)
💻 Local + CLI support → run servers yourself if you prefer
🎛 Remixable servers → browse, fork, and customize with ease
✨ What’s Next
Playground → test servers and tools directly in the app
Vibe-code tools → mix & match servers, tweak behaviour, create personal tools (think Lovable for MCPs)
What’s unique? We’re the first to make remote MCP servers truly plug-and-play so you can drop them into any client with zero friction. And because they’re open to the community, Disco becomes the one place to find, build, and share the best servers, with enough guardrails and good DX to keep quality high.
We’d love to hear: What’s the first tool you’d want your favourite agent/LLM client to use if spinning it up was dead simple?
💜 Guillaume & the StackOne team
PS: For anyone based in London, we'll be doing a Disco MCP hack evening with £5k Prize money: https://lu.ma/duwuc5hn
Disco.dev
@s1guillaume its been super fun working on this. Looking forward to seeing what integrations become popular!!
@s1guillaume Congrats on the launch! Does it auto-discover available servers or do I need to manually configure each MCP connection?
Disco.dev
@anwarlaksir Good question, once you create a Disco server, you do need to add it wherever you'd like to use it.
We'll look into having a way for all servers created to be auto-discovered but given the quantity of tools/MCP servers that most clients can deal with it's probably best to selectively link the servers you created to the various LLM clients that need them (eg. you might setup a Disco Jira MCP server with 4 tools to use within claude, while you won't care to connect it to let's say Cursor and you might want to connect another one, eg. Parallels instead to Cursor)
@s1guillaume MCP tool calling is powerful but the setup complexity has been holding back adoption. Making remote MCP servers truly plug-and-play is brilliant! Congrats on the launch!
Disco.dev
@kay_arkain Thanks Kay!
EnVsion AI
Amazing work StackOne team! I really like the smooth DX that makes adding the MCP server into one of the supported MCP clients so seamless. Very curious to try it out.
What are the next steps beyond the Hackathon?
Disco.dev
@ed_forson Thanks!!!!! The plan is to make the integrations super easy for AI to generate. Then expose orgs and teams so you can share your servers internally. It's what we have at StackOne and itss so good.
EnVsion AI
@matt_stackone Fantastic! I really think you guys are onto something here.
All the best!
Disco.dev
@ed_forson
First and foremost we're gonna add more integrations & tools but also feature so anyone can add tools without having to work on the open sourced repo but instead via a chat-based/lovable like experience, from prompt to new integration/tool available in your Disco account in < 30 seconds.
We'll also showcase Disco at the next AI Demo Day event we're doing (in Stockholm this time around) on the 3rd of September!
Asteroid
Congrats on the launch! I love what you've built and it's great to see it on producthunt. What's next on the integration roadmap?
Disco.dev
@joe_hewett Roadmap looks like this:
more integrations & tools
generate connectors via in app chat, from prompt to available in < 30 seconds.
private servers per user and sharing per organization (how we use it internally at StackOne)
hack next week with £5k prize money
and we'll also demo Disco at the next AI Demo Day in Stockholm on the 3rd of September!
As someone who's been wrestling with MCP servers for months, this is a breath of fresh air. The "zero infrastructure" promise actually delivers. You can spin up connections to Slack, Linear, and GitHub repos in under 5 minutes. No Docker containers, no environment configs, no deployment headaches.
I really like the "remix-ability" aspect. Being able to browse existing servers, fork them, and customize the tools for a specific use case is brilliant. It's like having a community-powered library of AI integrations that actually work out of the box. The fact that it's all open source (just starred the repo!) makes it even better because I can see exactly what's happening under the hood (and contribute back).
Quick question for the team: What about support for custom authentication, for example, for enterprise tools?
BTW, huge congrats on the launch!!!
Disco.dev
It’s been super fun to build this and we’ve loved using it internally over the last few months!!
Let us know what integrations we are missing :)
Baselime
Amazing product, I have needed the ability to get this level of control over my MCP servers for a while
Disco.dev
@thomas_ankcorn Great to hear - any MCP/Integration you'd like to have/use on a daily basis via Disco?
The Abacus
Congratulations @mattzcarey @s1guillaume and team on the launch!
Disco.dev
@s1guillaume @joonsang_lee Thanks!!! big love dude :)