AI agents (MCP)
Connect Claude, ChatGPT, Cursor or VS Code to your workspace. The agent builds guidance, checks it against your real product, shows you a preview and publishes when you say so.
Cueston runs a Model Context Protocol (MCP) server. Any AI app that speaks MCP can work in your workspace, with the same rules as the dashboard.
What an agent can do
| Step | What it covers |
|---|---|
| Understand your product | Workspace and rules, your app's pages, UI elements with verified selectors, events, saved audiences |
| Draft | Create and edit tours, tooltips, banners, checklists, surveys, announcements and segments, always as drafts |
| Check and preview | Validate against real pages, selectors and audience size; a preview link that shows the draft on your site |
| Ship | Publish or request review (your approval rules apply), unpublish, restore a version |
| Measure | Analytics, funnels, survey results, broken targets, "would this person see it?" |
Things to ask:
- "Announce the new scheduled reports feature: a What's new post and a tooltip on the Schedule button."
- "Build a five-step tour of the invoice flow for new admins and send me a preview link."
- "Which tours broke after last week's release? Repoint them and show me before you publish."
- "How did the onboarding checklist do in the last 30 days, and where do people drop off?"
Connect your AI app
Add Cueston as a connector with this URL:
https://app.cueston.com/api/mcpThe app opens Cueston so you sign in, pick the workspace and choose what it may do. There are no keys to copy.
- In Claude (web or desktop), open Settings → Connectors → Add custom connector.
- Name:
Cueston. URL:https://app.cueston.com/api/mcp. - Choose Connect and sign in to Cueston.
- In ChatGPT, open Settings → Connectors. Turn on Developer mode if it's offered.
- Add a connector. Name:
Cueston. URL:https://app.cueston.com/api/mcp. Authentication: OAuth. - Sign in to Cueston.
claude mcp add --transport http cueston https://app.cueston.com/api/mcpThen run /mcp in Claude Code and choose Authenticate.
Cursor asks you to sign in on first use.
{
"mcpServers": {
"cueston": { "url": "https://app.cueston.com/api/mcp" }
}
}VS Code asks you to sign in on first use.
{
"servers": {
"cueston": { "type": "http", "url": "https://app.cueston.com/api/mcp" }
}
}Other MCP apps work the same way: give them the server URL and they find the sign-in page on their own. You need a Cueston account in the workspace; create one if you don't have one yet.
For scripts, CI and apps that can't sign in, use an API key instead.
Let the agent set it up
Coding agents such as Claude Code, Cursor, Codex or Windsurf can add Cueston to their own settings. Paste this into the chat:
Add the Cueston MCP server to your MCP settings.
Name: cueston
URL: https://app.cueston.com/api/mcp (Streamable HTTP)
Sign-in: OAuth. There's no API key to copy;
I'll sign in to Cueston in my browser.
If you can only run local (stdio) servers, use the npm package instead:
command "npx", args ["-y", "cueston", "mcp"].
I'll give you an API key for CUESTON_API_KEY.
When it's added, tell me how to finish signing in.
Then call get_workspace and tell me what you see.Apps that only run local servers
The cueston package on npm runs Cueston as a local (stdio) MCP server. It forwards every request to the hosted server with your API key:
{
"mcpServers": {
"cueston": {
"command": "npx",
"args": ["-y", "cueston", "mcp"],
"env": { "CUESTON_API_KEY": "sk_…" }
}
}
}Leave out env if you've run cueston login on this computer. The same package is the Cueston CLI.
Your first request
Once you're connected, paste this to see the whole loop on your own product:
Use the Cueston tools.
1. Look at my workspace, the pages Cueston knows
and the guidance that's already live.
2. Suggest three pieces of guidance that would help
new users most, and say why.
3. Build the first one as a draft, with selectors
you've checked on the real page.
4. Validate it and send me a preview link.
Don't publish until I say so.Apps that support MCP prompts also list ready-made workflows in their prompt or slash-command menu: see Prompts.
Next
- Access and safety: what an agent can and can't do.
- Tools and prompts: the full catalog.
- Reference: transport, sign-in and limits, for client builders.