1What your agent can do
- Screen names — up to 10 at a time: exact-match USPTO trademark search (with the classes that hold a live mark), five domains, a Delaware entity check and the free Brand Score.
- Open a full report — trademark records, entity records in five states, social handles on eight platforms, search competition and an AI analysis.
- List your reports — so it doesn't re-check names you already have.
2Create a token
In Settings → Agents & API, name a token (for example “Claude Desktop”) and copy it. It's shown once. Treat it like a password: it lets the agent act on your account. You can revoke it any time.
3Connect your agent
The server address is https://www.brandscreen.com/mcp. Replace YOUR_TOKEN below with your token.
Claude Code
claude mcp add --transport http brandscreen https://www.brandscreen.com/mcp \
--header "Authorization: Bearer YOUR_TOKEN"Cursor — add to .cursor/mcp.json
{
"mcpServers": {
"brandscreen": {
"url": "https://www.brandscreen.com/mcp",
"headers": { "Authorization": "Bearer YOUR_TOKEN" }
}
}
}Claude Desktop — add to claude_desktop_config.json (needs Node.js)
{
"mcpServers": {
"brandscreen": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://www.brandscreen.com/mcp", "--header", "Authorization:${AUTH}"],
"env": { "AUTH": "Bearer YOUR_TOKEN" }
}
}
}Any other client that supports remote MCP servers over HTTP works the same way: the server address plus an Authorization: Bearer header. One-click sign-in (no token) is coming later.
4Limits
- Screening: 20 names a day on the Free plan, 100 on paid plans.
- Full reports: 10 new reports a day, shared with the website. Re-opening a report you already have is free.
5What the results mean
BrandScreen is a screen, not legal clearance. The trademark search matches the exact name, so similar marks can still matter, and a clean result means nothing matching was found — not that the name is safe to use. Your agent receives these caveats with every result. For a filing decision, talk to a trademark attorney.