Your AI agent might
already violate EU law.
TSOT flags EU AI Act violations as you build — inside Cursor, Claude, and Windsurf. No API key. No account. Free forever.
What it looks like in Cursor
One prompt. Instant statutory citation. No browser tab required.
This agent generates synthetic persona responses without an explicit AI disclosure. Under Reg. (EU) 2024/1689 you must notify users they are interacting with an AI system at the outset of every interaction.
Required action
Add a disclosure notice at session start — backed by [SOT-COMP-2026] and statutory citation [Art. 50(1)].
No other compliance tool is backed by 9,200+ peer-reviewed papers.
Every audit cites real research — empirical HCI findings mapped to the exact statutory article that applies to your system. Not AI-generated advice. Not legal opinion.
Showing 3 of 9,200+ indexed records · Updated continuously via OpenAlex
Add it once. It runs inside
your editor forever.
Add the MCP server
Add npx -y tsot-mcp-server to your Cursor, Claude Desktop, or Windsurf MCP config file.
Ask while you build
Prompt your agent: "Audit this prompt for EU AI Act risk" — no context switching, no browser tab.
Get citations, not opinions
Every response includes a statutory article tag like [Art. 50] and a citable HCI research paper reference from the Ledger.
Free. Local.
No keys.
Never ship an EU AI Act violation again.