SnapStack

Capture the visible browser tab in one click and stack it locally for any MCP-capable LLM client.

Hand your browser screenshots to your AI — no copy-paste, no uploads. SnapStack captures the visible browser tab (or a rectangular area you select) in one click and stacks it on your own machine. Your AI assistant then reads the screenshots on demand through the Model Context Protocol (MCP) — so it works with any MCP-capable client, such as Claude Code. How it works: 1. Click the SnapStack button (or "Capture area") — the shot is stacked locally. 2. The toolbar badge shows how many captures are waiting. 3. Ask your LLM client to look at your screenshots; it reads them over MCP. SnapStack pairs with a small companion server you run on your own computer (open source, installed separately). The extension talks only to that local server at 127.0.0.1 — nothing is ever uploaded. Features: one-click tab capture · area capture · full-page capture · local numbered stack + live badge · manage from the popup (copy path, copy all, delete, open folder) · WebP/PNG/JPEG with optional width downscale · keyboard shortcuts · settings shared across browsers. 100% local & private — no account, no telemetry, no remote server. Captures go only to 127.0.0.1, in a folder you control. Requirements — the SnapStack companion server (free, open source) + an MCP-capable LLM client. Open source — github.com/bgaze/snapstack-extension