Two ways to give your AI structured perception of the web. The extension is for hands-on capture; E2LLM MCP connects your AI tools to your browser directly. Both speak the same SiFR representation.
Pick any element and get a structured SiFR snapshot — straight to your clipboard or a file. Paste it into any LLM chat to debug, review, or audit. SiFR generation runs in your browser, the page DOM never leaves your machine, and no account is required.
Connect your AI tools to your browser over the Model Context Protocol. Your assistant requests a live SiFR snapshot, then clicks, fills, and scrolls — eyes and hands on the page you are already signed in to.
● Store listing rolling outmcp.e2llm.com. The companion MCP browser extension is rolling out to the Chrome and Firefox stores — the setup guide walks you through getting connected today.E2LLM captures structured state from supported standard DOM surfaces. SiFR is a model-readable representation of rendered page state. Free MCP and paid tiers are listed on the pricing page.
Capture a snapshot when you want one and paste it wherever you like. Best for debugging, code review, accessibility checks, and one-off questions where you stay in control of each capture.
Your assistant reads and acts on the live page through your browser. Best for autonomous workflows, multi-step tasks, and QA passes where the AI needs to both see and act without copy-paste.
Claude · Claude Code · ChatGPT · Gemini · Grok · Cursor · VS Code · OpenAI Codex — and any client that speaks the Model Context Protocol.
Prompts & examples on GitHub →