Add Dotsy to your AI.One URL. No setup.
Paste this link into Claude, ChatGPT, Cursor or any MCP client. No API keys, nothing to install.
https://mcp.dotsy.ai/mcp
Then pick your AI below. Two minutes, tops.
Pick your AI.Three steps or fewer.
Claude
Web and desktop apps, any plan. Free includes one custom connector.
- Open Customize → Connectors (Settings → Connectors in older versions).
- Click +, then Add custom connector. Name it Dotsy, paste the URL, click Add.
- Click Connect and sign in to Dotsy in the tab that opens. Done.
Team or Enterprise? An owner adds it once under Organization settings → Connectors, then everyone clicks Connect.
ChatGPT
Through Developer mode, a beta on the web for Plus, Pro, Business, Enterprise and Edu.
- In Settings, turn on Developer mode (under Security and login, or Apps → Advanced).
- Create a new app: name it Dotsy, paste the URL, set authentication to OAuth, and confirm you trust it.
- Sign in to Dotsy, then turn Dotsy on from the Developer mode tool in any chat.
ChatGPT marks custom apps as unverified and asks before write actions. That is expected. Menu names move around in this beta; the URL never does.
Cursor
One click, then sign in.
- Click Add to Cursor and confirm Install in Cursor.
- In Cursor Settings → Tools & MCP, click Connect next to dotsy.
- Sign in to Dotsy in your browser. Cursor picks it up.
~/.cursor/mcp.json
{
"mcpServers": {
"dotsy": { "url": "https://mcp.dotsy.ai/mcp" }
}
}
VS Code
For Copilot Chat in Agent mode.
- Click Add to VS Code and choose Install.
- When VS Code asks to authenticate, click Allow and sign in to Dotsy.
- Open Chat in Agent mode. Dotsy's tools are in the tools list.
.vscode/mcp.json
{
"servers": {
"dotsy": { "type": "http", "url": "https://mcp.dotsy.ai/mcp" }
}
}
Claude Code
One command in your terminal.
- Run the command below.
- In Claude Code, type
/mcp, pick dotsy, then Authenticate. - Sign in in the browser tab that opens. That's it.
claude mcp add --transport http dotsy https://mcp.dotsy.ai/mcp
Add --scope user to use Dotsy in every project. Big site? Claude Code can push it from disk with the Dotsy CLI.
$ claude mcp add --transport http dotsy \
https://mcp.dotsy.ai/mcp
Added HTTP MCP server dotsy
> /mcp
❯ dotsy needs authentication
Authenticate · opening browser…
✔ dotsy connected · 40+ tools
Devin Desktop formerly Windsurf
Same editor, new name. Same steps.
- In the agent panel, open the … menu → MCP → Open MCP config file.
- Add the Dotsy entry below and save.
- Refresh MCP servers and sign in to Dotsy when the browser opens.
{
"mcpServers": {
"dotsy": { "serverUrl": "https://mcp.dotsy.ai/mcp" }
}
}
Note the key: Devin Desktop uses serverUrl, not url. Still on an older Windsurf build? The steps are the same.
Any MCP client
Dotsy speaks Streamable HTTP with OAuth 2.1 sign-in.
- Add a remote (HTTP) MCP server with the Dotsy URL.
- Sign in when your client opens the browser.
- No remote URL support? Run it through
mcp-remoteas a command instead.
{
"mcpServers": {
"dotsy": { "url": "https://mcp.dotsy.ai/mcp" }
}
}
Scripts and CI: create an API key in the dashboard and send it as a Bearer header instead of signing in.
{
"mcpServers": {
"dotsy": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://mcp.dotsy.ai/mcp"]
}
}
}
Prefer a terminal?
The Dotsy CLI pushes a folder straight from your disk. Made for big sites, CI, and AI agents that run commands.
Needs Node.js 20 or later. Leave off --publish to check a draft first.
- Install it.
npm i -g @dotsy-ai/cli
- Log in once. Paste an API key from your dashboard.
dotsy login
- Push a folder. New site name? It offers to create it.
dotsy push my-site ./site --publish
Building an integration?
Use the REST API, the same one the dashboard, the CLI and our MCP server run on. Create sites, write pages, upload files, read form entries and publish.
Same API key as the CLI. API reference · OpenAPI spec
curl https://api.dotsy.ai/api/v1/sites \
-H "Authorization: Bearer $DOTSY_API_KEY" \
-H "Content-Type: application/json" \
-d '{"slug":"my-site"}'
Ideas
- Today’s specials, straight from your till
- A page for every new product
- Form entries into your CRM
- A closure banner on every page
- A site for each of your customers
Then just ask.Your AI builds the draft. Sign in when you publish.
start_siteNo account neededput_content4 pagesedit_pageSchedulepublish_siteLiveSlow down.See this week's classes
- Mon7:00Morning VinyasaBook
- Tue18:30Yin & BreathBook
- Thu12:15Lunch FlowBook
Some AI apps ask you to sign in when you first connect.
Questions
Does Dotsy have its own AI?
No. You use the AI you already have — Claude, ChatGPT, Cursor or any MCP client. Dotsy gives it the tools to build, edit and publish your site. You never pay us for AI.
What is MCP?
The Model Context Protocol is an open standard that lets an AI assistant use outside tools. Dotsy's MCP server gives your AI the tools to build, edit and publish your site. It can start a new draft right away; it publishes, or touches your existing sites, only after you sign in.
Which AI tools work?
Anything that connects to a remote MCP server: Claude on the web, desktop and in Claude Code, ChatGPT with Developer mode, Cursor, VS Code with Copilot, Devin Desktop (formerly Windsurf), and many more. Clients that only run local servers can use npx mcp-remote.
Can I use Dotsy without an AI?
Yes. Everything you can ask your AI to do also works in the Dotsy dashboard. Prefer a terminal? The Dotsy CLI pushes a folder from your disk.
Is there an API?
Yes. A REST API with an OpenAPI spec lets you create sites, write pages, upload files, read form entries and publish from your own code. It uses the same API key as the CLI. See how it works.
Do I pay for AI?
No. You use the AI plan you already have, and Dotsy never charges AI credits. You only pay Dotsy for hosting, and that starts at $0.