Element to LLM MCP 作者： insitu.im

E2LLM MCP gives your AI eyes and hands in a real browser — structured perception plus action, working with the AI you already use.

WHAT IT DOES
This extension connects your browser to your AI assistant (Claude, ChatGPT, Grok, Perplexity, and other MCP-capable clients) through the e2llm relay (mcp.e2llm.com). Your AI can then:
• See pages as structured, deterministic data (SiFR format) instead of screenshots — visible text, forms, page structure
• Act in the browser on your behalf: click, type, fill forms, navigate — always explicit, always under your control
• Work across tabs with per-session grants you can revoke at any time

HOW IT WORKS
1. Install the extension and sign in at account.e2llm.com (free tier available)
2. Add the E2LLM connector to your AI client (mcp.e2llm.com) — setup guide: e2llm.com/docs/setup
3. Ask your AI to read or act on any page — the extension is the bridge, your AI is the driver

NOTE: this is the MCP bridge extension — a different product from our standalone "Element to LLM" capture extension. The standalone works fully locally (copy page context into any chat; no account, no network). THIS extension connects to the e2llm relay service and requires an account. If you only need local capture, the standalone extension is what you want.

PRIVACY
Sessions are explicit and grant-based. Sensitive fields (passwords, card numbers, one-time codes) are redacted before storage; session data is purged after 30 days. Full policy: e2llm.com/privacy

Requires: an e2llm account and an MCP-capable AI client.