{
  "$type": "site.standard.document",
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreih26hhwipdw2nez52qrnzy7zkkmbornutcjx6hs4rpkspdgp6jrtm"
    },
    "mimeType": "image/png",
    "size": 95443
  },
  "description": "Bring Val Town to your favorite LLM",
  "path": "/mcp",
  "publishedAt": "2025-11-14T00:00:00.000Z",
  "site": "at://did:plc:a2rdzfdxkjwerrfrpbwcipb2/site.standard.publication/3jd443afc2222",
  "textContent": "On Val Town, you deploy JavaScript in 100ms. Now with the Val Town MCP server,\nyou can do that from Claude, ChatGPT, Cursor, VSCode, or wherever you do your AI\ncoding.\n\n\n\nIf you've been following my tweets recently –\n\"I've gotta rant about LLMs, MCP, and tool-calling for a second\",\n\"MCP is mostly nonsense\",\n\"MCP is overhyped\" – you\nmight be surprised by this announcement. Well, how did you think I got those\nsalty takes except by building an MCP server?\n\nYes, I think MCP is dubious as a protocol. But for now, MCP is the right way for\nVal Town to meet developers where they are. In Cursor or Claude Code or Zed or\nwherever. For example, here we use Claude Code to make a blog. Every edit is\nimmediately live and deployed on Val Town.\n\n<p style=\"position: relative; aspect-ratio: 16/9; border: 1.5px solid grey; border-radius: 10px; overflow: hidden;\">\n  <iframe\n    src=\"https://player.mux.com/jpAEUVeTm7b500YSTNSSueYEG2EM02YNirJiPePOAow02s?preload=metadata&muted=true\"\n    style=\"position: absolute; top: 0; left: 0; width: 100%; height: 100%; border: 0;\"\n    allow=\"accelerometer; gyroscope; encrypted-media; picture-in-picture\"\n    allowFullscreen=\"true\"\n  ></iframe>\n</p>\n\nWe have guides for some of the popular LLMs:\n\n- Claude Code\n- ChatGPT\n- Claude Web/Desktop\n\nBut the Val Town MCP server should work with any MCP client. If you'd like a\nhand with setup, ask in our discord server or send\nus an email.\n\nWhy MCP\n\nMCP is not perfect (again, see tweets), but it has a few things going for it:\n\n1. Cheaper – Don't pay us for credits. Pay your inference provider\n   directly.\n2. Better – Use whatever state-of-the-art LLM you want. We at Val Town don't\n   have to fast-follow it.\n3. Val Town everywhere – Get the best parts of Val Town – instant\n   deployments, built-in SQLite, etc – in your favorite LLM coding tool.\n\nMCP also allows us to\nship faster. Traditional\nAPIs require careful versioning to prevent breaking changes, but an MCP server\ncan change continuously because LLMs read the spec and run inference at\nruntime.\n\nFast feedback loops\n\nThere's a common thread running through every feature we build – AI or\notherwise: enabling fast feedback loops.\n\n> Creators need an immediate connection to what they're creating.\n>\n> If you make a change, you need to see the effect of that immediately.\n>\n> \\- Bret Victor, Inventing on Principle\n\nWhen you – or your LLM – make an edit on Val Town, your code is deployed in\n100ms. This allows you to have insanely fast feedback loops in your production\nenvironment. No need to wait a minute or two to see how it'll actually look\nwhen deployed. Every change is immediately live, at a public URL.\n\nVal Town isn't an AI company – we're a developer tools company – but this\nalways-deployed model works quite well with LLMs. Just give your favorite LLM a\nbranch, and the code it writes will be alive and sharable by default.\n\nBring Val Town MCP to your favorite LLM,\nand let us know what you think.",
  "title": "Introducing Val Town MCP"
}