Capture any browser tab in one click and stack it locally — your AI reads the screenshots on demand.

snapstack lets you hand browser screenshots to your AI assistant without copy-pasting images. Click the extension, your shot is stacked locally; ask your LLM to "look at my screenshots" and it picks them up — through the Model Context Protocol (MCP), so it works with any MCP-capable client (Claude Code, and others).

100% local. Captures go only to a small server on your own machine (127.0.0.1). Nothing is ever uploaded, no account, no telemetry. See PRIVACY.md.

snapstack has two halves: this extension (capture) and a small always-on local server (snapstack-server) that holds the stack and serves it to your LLM. Both install in a couple of minutes — see Installation.

How it works

[extension]  --click → capture-->  ┐
                                    ▼
                      [local server]   127.0.0.1
                         ├─ stores ─►  a folder on your disk
                         └─ MCP    ◄──  your LLM client asks for the screenshots

1. You browse and capture the screens you care about — they pile up in a local stack.
2. Your LLM client calls snapstack's MCP tools and gets a list of the pending shots.
3. It reads the ones it needs, straight from your disk. The captures stay until you clear them.

Using it

Click the snapstack icon to open the dropdown:

• Capture — take a shot of the current tab and add it to the stack. The icon badge shows how many are stacked.
• Capture area — drag a rectangle over the page and stack just that region; press Esc to cancel. (Not available on browser-internal pages such as chrome:// or the web store.)
• Grid — your captures, two per row, each tagged with its number (the handle your LLM uses). Hover for Delete / Copy path; click a shot to open it full-size.
• Toolbar — Delete all, Open folder (reveal the stack in your file manager), Copy all paths.

A typical session: capture a few screens → tell your AI "have a look at my screenshots" → it reads them and answers. When you're done, clear them (from the dropdown, or let the AI do it).

The dropdown follows your browser's language (English, French, Spanish, German, Italian, Japanese, Portuguese-BR, Russian, Simplified Chinese), falling back to English.

Exposed MCP tools

Your LLM client uses these three tools — no image bytes are ever pushed to the model; it gets a lightweight list and reads only the files it wants.

Tool	What it does
get_screenshots	Lists the pending captures (number, file path, size, page URL/title) — read-only, never deletes. Pass numbers (e.g. [1,3]) for specific ones.
clear_screenshots	Removes captures from the stack — specific numbers, or all of them. The only destructive tool.
count_screenshots	Just how many captures are waiting.

Installation

Three steps: the server, your MCP client, this extension.

1. Install the local server

Needs Node.js ≥ 18 and git. Check with node -v and git --version.

It's a permanent background tool, so it installs into your system's standard app location (not Downloads). Pick your OS — copy the whole block and paste it into a terminal:

macOS

# 1. Clone the server into the macOS app-support directory
git clone https://github.com/bgaze/snapstack-server.git "$HOME/Library/Application Support/snapstack"
cd "$HOME/Library/Application Support/snapstack"

# 2. Install dependencies (just two small packages)
npm install

# 3. Start it now + at every login, with crash-restart and auto-update (launchd)
./deploy/install-macos.sh

Linux

# 1. Clone the server into the standard data directory
git clone https://github.com/bgaze/snapstack-server.git "$HOME/.local/share/snapstack"
cd "$HOME/.local/share/snapstack"

# 2. Install dependencies (just two small packages)
npm install

# 3. Start it now + at every login, with restart and auto-update (systemd --user)
./deploy/install-linux.sh

Windows (PowerShell)

# 1. Clone the server into your local app data
git clone https://github.com/bgaze/snapstack-server.git "$env:LOCALAPPDATA\snapstack"
Set-Location "$env:LOCALAPPDATA\snapstack"

# 2. Install dependencies (just two small packages)
npm install

# 3. Start it now + at every login, with auto-update (scheduled task)
.\deploy\install-windows.ps1

That's it — the server now starts on login and updates itself (git pull) each time it launches. To run it just once instead, skip step 3 and use npm start.

2. Register snapstack with your MCP client

The server speaks MCP over HTTP at http://127.0.0.1:4123/mcp (it must be running first).

Claude Code — one command:

claude mcp add --transport http --scope user snapstack http://127.0.0.1:4123/mcp

Other clients — add an HTTP MCP server; most accept this shape (consult your client's docs for the exact syntax):

{
  "mcpServers": {
    "snapstack": { "type": "http", "url": "http://127.0.0.1:4123/mcp" }
  }
}

3. Install the extension

Browser	Install
Chrome	Chrome Web Store
Edge	Edge Add-ons (or install from the Chrome Web Store)
Firefox	Firefox Add-ons

Developer install (unpacked, from source)

• Chrome / Edge: chrome://extensions → enable Developer mode → Load unpacked → select this folder. (A harmless "Unrecognized manifest key 'background.scripts'" warning is expected — that key is for Firefox.)
• Firefox: about:debugging#/runtime/this-firefox → Load Temporary Add-on → select manifest.json.

Use it with Claude

Two touches make snapstack feel native in Claude Code.

Skip the per-call confirmation — allow the tools up front. Add to ~/.claude/settings.json (covers every project):

{
  "permissions": {
    "allow": [
      "mcp__snapstack__count_screenshots",
      "mcp__snapstack__get_screenshots",
      "mcp__snapstack__clear_screenshots"
    ]
  }
}

get_screenshots and count_screenshots are read-only; clear_screenshots is the only one that deletes — leave it out of the list if you'd rather confirm deletions by hand.

Add a snap shortcut — drop this in your ~/.claude/CLAUDE.md so a single word pulls in your captures:

## snap shortcut
When I type `snap` (optionally with capture numbers and/or an instruction):
call `get_screenshots` to fetch the pending captures, then act on them in
context. `snap clear` (or `snap clear 01 02`) → `clear_screenshots`.

Now you just capture in the browser and type snap — Claude reads exactly what's on your screen.

Troubleshooting

• "Capture server not started" — the local server isn't running. Start it (or check its auto-start); see snapstack-server.
• Red ! badge — the extension can't reach the server. Make sure it's running on 127.0.0.1:4123.
• Captures saved as PNG — your browser can't encode WebP, so snapstack falls back to PNG automatically. Normal.

Advanced settings

The defaults work out of the box. Power users can tweak these via the extension's storage.local: serverBase (http://127.0.0.1:4123), format (webp/png), quality (0.85), maxEdge (1568 px, 0 = no downscale).

Packaging for the stores

The committed manifest.json carries both background.service_worker (Chrome) and background.scripts (Firefox) so it loads unpacked in either browser. A store package must carry only its own browser's key, so build per-browser bundles:

npm install      # dev-only: eslint
npm run build    # → dist/chrome/ + dist/firefox/ and a zip for each

dist/snapstack-chrome-<version>.zip goes to the Chrome Web Store / Edge Add-ons; dist/snapstack-firefox-<version>.zip to Firefox AMO. Requires the system zip.

Privacy & License

No data collection, nothing leaves 127.0.0.1. See PRIVACY.md. MIT — see LICENSE.