Connect Klevo to Claude, ChatGPT and Cursor

Klevo is an MCP server. You add it once to the AI you use, and from then on “record a demo of my app” comes back as an MP4 in the chat. This guide walks through each app.

Claude.ai, desktop and mobile

Claude signs in to your Klevo account; there is no key to paste.

1. Create a free account at klevo.si and set your password from the welcome email.
2. In Claude, open Settings → Connectors and choose Add custom connector.
3. Name it Klevo and paste https://klevo.si/mcp as the URL.
4. Click Connect, sign in to Klevo and choose Allow.

The connector follows your account, so it works in the desktop and mobile apps too. Nothing runs on your machine: the browser that records your app lives on Klevo’s servers.

ChatGPT

1. Open Settings → Connectors (on some plans, developer mode has to be on).
2. Add a custom MCP server named Klevo with the URL https://klevo.si/mcp and authentication set to OAuth.
3. Sign in to Klevo and choose Allow.

Claude Code

Create a key on the Connect page, then run once:

claude mcp add --transport http klevo https://klevo.si/mcp --header "Authorization: Bearer klv_live_…"

Cursor and VS Code

Add this to ~/.cursor/mcp.json (Cursor) or .vscode/mcp.json (VS Code), with the key from the Connect page:

{ "mcpServers": { "klevo": { "url": "https://klevo.si/mcp", "headers": { "Authorization": "Bearer klv_live_…" } } } }

If your app has a login

Klevo needs a way in before it can record anything behind a login. Save one on the Secrets page, or ask your AI to do it:

Use Klevo’s auth tool to inspect https://staging.myapp.com/login and tell me how to save a login secret.

For a plain form, a test account saved as a password secret is enough. For Google sign-in, two-factor or a captcha, export a browser session and save it as a storage state. The recording guide covers both.

The first demo

Start a new chat and ask:

Record a demo with Klevo of creating a project on https://staging.myapp.com/projects: click New project, name it “Q4 launch”, save it. Desktop, 16:9, browser frame, with captions. Log in with DEMO_LOGIN.

The AI opens the page, clicks and types through the flow while checking a screenshot after each step, writes a caption per step and finishes. The video renders in the background; ask “is the video ready?” after a minute, or open Videos.

Want the same demo for a story? Say “render it 9:16 on a phone frame”. The script is saved, so there is nothing to record again.

Every video is kept in your account under Videos, every saved login under Secrets, and every call shows up under Usage.