MCP server — drive Clearcote from an AI agent
Point Claude Desktop, Cursor or Cline at the Clearcote MCP server and let the model drive one shared stealth browser through ~20 tools. The persona is set once via env, so the tool surface stays clean — the agent works, the identity stays coherent underneath.
Configure your MCP client
Add Clearcote to your client's MCP config (Claude Desktop shown; Cursor/Cline use the same shape):
{
"mcpServers": {
"clearcote": {
"command": "npx",
"args": ["-y", "clearcote-mcp"],
"env": {
"CLEARCOTE_FINGERPRINT": "acct-1",
"CLEARCOTE_PLATFORM": "windows",
"CLEARCOTE_LICENSE_KEY": "cc_lic_... (optional — selects the Pro build)"
}
}
}
}Prefer Python? The server is on PyPI too:
pip install clearcote-mcp # or: npx -y clearcote-mcpTools
A single shared browser is driven by around twenty tools. The most-used:
| Tool | What it does |
|---|---|
read_page | Read the live page as text/markdown |
page_elements | List interactive elements with stable refs |
click | Click an element |
fill_field | Type into an input / textarea |
screenshot | Capture the viewport or full page |
navigate | Go to a URL / back / forward / reload |
save_profile | Persist the current persona/profile |
get_cdp_endpoint | Return the CDP URL for a raw client |
…plus scrolling, waiting, tab management, cookie/localStorage control, and profile management. The agent reads the page, decides, and acts — all through one persona-consistent Clearcote instance.
Persona via environment
Everything you'd pass to the fingerprint flags is available as aCLEARCOTE_* env var, so the model never has to think about identity:
CLEARCOTE_FINGERPRINT=acct-1 # seed -> stable identity
CLEARCOTE_PLATFORM=windows # windows | linux | macos | android
CLEARCOTE_BRAND=Edge # Chrome (default) | Edge
CLEARCOTE_ACCEPT_LANGUAGE=en-US
CLEARCOTE_TIMEZONE=America/New_York
CLEARCOTE_LICENSE_KEY=cc_lic_... # optional -> license-gated Pro buildThe MCP server drives the browser via CDP, sonavigator.webdriverstaysfalseand the engine'sRuntime.enableneutralization keeps the attached client invisible to the page. Two ways to automate: an LLM outside the browser (this MCP server) or the in-browser AI agent that runs inside the process.
Need a raw endpoint instead of MCP? See Deployment & Docker.