@silbercue/chrome

v1.3.0

Published

4 days ago

SilbercueChrome MCP Server — Chrome browser automation via CDP

Downloads

2,314

0High
0Medium
0Low

silbercue

SilbercueChrome

The most token-efficient MCP server for Chrome browser automation. Direct CDP, a11y-tree refs, multi-tab ready — 1670+ TypeScript tests, 235+ Python tests.

Built for Claude Code, Cursor, and any MCP-compatible client.

Looking for an alternative to Playwright MCP, Browser MCP, or claude-in-chrome? SilbercueChrome talks to Chrome directly via the DevTools Protocol — no Playwright dependency, no Chrome extension bridge, no single-tab limit. One command to install, zero config. See benchmark comparison below.

Why SilbercueChrome?

Every Chrome MCP server has the same problem: too many tokens, too few reliable refs. Screenshots eat 10-30x more tokens than text trees. Selector-based refs break the second the DOM rerenders. Extension bridges (Browser MCP) get stuck on the connected tab. Playwright wrappers spin up a new browser instance for every session.

SilbercueChrome fixes this. It talks directly to Chrome via CDP (same protocol Playwright and Puppeteer use internally), returns an accessibility-tree-based reference map, and caches it across calls so click(ref: 'e5') and type(ref: 'e7', ...) survive scrolls and DOM updates.

| What you get | Playwright MCP | Browser MCP | claude-in-chrome | browser-use | SilbercueChrome | |---|---|---|---|---|---| | Hardest benchmark (35 tests, LLM-driven) | 29/31 (563s) | cannot finish | (pending re-bench) | (pending re-bench) | 30/31 Free: 598s | | ∅ Tool-Response (Tokens est.) | 362 | — | — | — | 201 (1.8× smaller) | | P95 Tool-Response (Chars) | 8.068 | — | — | — | 2.328 (3.5× smaller) | | view_page avg (Chars) | 6.084 (browser_snapshot) | — | — | — | 1.124 (5.4× smaller) | | Multi-tab support | Yes | No (single tab) | Yes | Partial | Yes | | Connection | New browser | Extension bridge | Extension | Subprocess | Direct CDP (pipe or WebSocket) | | Ref system | Playwright refs | Playwright refs | CSS selectors | Screenshots | A11y-tree refs (stable across DOM changes) | | Drag & drop | Yes | No | Partial | No | Yes (native CDP mouse events) | | Shadow DOM + iframe | Yes | Yes | Partial | No | Yes (with OOPIF session support) | | Multi-step plan execution | — | — | — | — | run_plan — server-side plan executor with variables, conditions, suspend/resume |

Quick Start

Install in Claude Code

One command — installs globally for all projects:

claude mcp add --scope user silbercuechrome npx -y @silbercue/chrome@latest

Important: after claude mcp add you must fully quit and reopen Claude Code. /mcp reconnect is not enough — Claude Code reads the mcpServers config only at session start and caches it. After the restart, the first tool call auto-launches Chrome visible (no headless, no port setup). Done.

To enable parallel Python Script API access, add --script to the args: claude mcp add --scope user silbercuechrome npx -y @silbercue/chrome@latest -- --script

Install in Cursor

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "silbercuechrome": {
      "command": "npx",
      "args": ["-y", "@silbercue/chrome@latest"]
    }
  }
}

For parallel Python Script API access, use "args": ["-y", "@silbercue/chrome@latest", "--", "--script"]

Install in Cline

Add to your cline_mcp_settings.json:

{
  "mcpServers": {
    "silbercuechrome": {
      "command": "npx",
      "args": ["-y", "@silbercue/chrome@latest"]
    }
  }
}

Install in other MCP clients

Any client that supports stdio MCP servers: npx -y @silbercue/chrome@latest with no arguments.

Try it — your first prompt

After installing, paste this into your AI coding assistant:

Open mcp-test.second-truth.com, read the page, and fill the contact form with Name "Test User" and Email "[email protected]".

This exercises three core tools in sequence: navigate loads the page, view_page reads the accessibility tree with stable element refs, and fill_form fills multiple fields in one call. You should see Chrome open, the page load, and the form filled — all without writing a single line of code.

Install Pro via Homebrew

Pro adds virtual_desk, switch_tab, dom_snapshot, parallel tabs in run_plan, ambient context hooks, and an operator hook pipeline on top of the 18 Free tools. Three commands, no JSON edits:

brew install silbercue/silbercue/silbercuechrome
claude mcp add --scope user silbercuechrome /opt/homebrew/bin/silbercuechrome
silbercuechrome activate SCC-XXXX-XXXX-XXXX-XXXX

Important — restart Claude Code completely after claude mcp add. /mcp reconnect is not enough. Claude Code reads the mcpServers config only at session start and caches it; the old command is re-used even after reconnect. Fully quit Claude Code and reopen it so the new silbercuechrome server is picked up.

After the restart, silbercuechrome status should print Tier: Pro, Tools: 23. Get a license at polar.sh/silbercuechrome — the key arrives by email and can be activated as shown above.

Google Chrome must be installed on the machine — SilbercueChrome auto-launches Chrome via CDP at runtime, but it does not install Chrome for you.

Upgrading an existing install:

brew update && brew upgrade silbercue/silbercue/silbercuechrome

Your license key stays in ~/.silbercuechrome/license-cache.json and survives the upgrade — no re-activation needed.

Uninstall

# Free (npx install)
claude mcp remove --scope user silbercuechrome

# Pro (Homebrew install)
silbercuechrome deactivate   # wipes license cache
claude mcp remove --scope user silbercuechrome
brew uninstall silbercue/silbercue/silbercuechrome

Script API (Python)

A third way to use SilbercueChrome — deterministic browser automation from Python, without an LLM in the loop. Scripts use the same tool implementations as the MCP server (Shared Core) — every improvement to click, navigate, fill_form etc. automatically benefits your scripts too. The MCP server handles AI-driven workflows; the Script API is for repeatable scripts you write yourself.

Installation

pip install silbercuechrome

That's it. Chrome.connect() auto-starts the SilbercueChrome server as a subprocess — no manual Chrome launch or port setup needed.

Legacy single-file alternative: For quick prototyping you can copy python/silbercuechrome.py into your project. This uses v1 direct CDP and does not benefit from server-side improvements — use pip install for the full Shared Core experience.

How it works

Python Script                        Escape Hatch (Power User)
    │                                    │
    ▼                                    ▼
HTTP POST /tool/{name}              WebSocket (CDP)
Port 9223                           Port 9222
    │                                    │
    ▼                                    │
SilbercueChrome Server                   │
    │                                    │
    ▼                                    │
registry.executeTool()                   │
    │                                    │
    ▼                                    │
Tool Handler                             │
(click.ts, navigate.ts, ...)             │
    │                                    │
    ▼                                    ▼
Chrome ◄─────────── CDP ────────────────►

Your script sends HTTP requests to the SilbercueChrome server on port 9223. The server executes the exact same tool handlers that the MCP server uses — one codebase, one test suite (1670+ tests), two access paths.

Auto-Start

Chrome.connect() finds and starts the server automatically:

Running server — checks if port 9223 already responds, connects immediately
PATH binary — finds silbercuechrome in PATH (e.g. via Homebrew), starts it with --script
npx fallback — runs npx -y @silbercue/chrome@latest -- --script
Explicit path — Chrome.connect(server_path="/path/to/silbercuechrome") for custom setups

Example: Login + Data Extraction

from silbercuechrome import Chrome

chrome = Chrome.connect()

with chrome.new_page() as page:
    page.navigate("https://competitor.example.com/login")
    page.fill({"#email": "[email protected]", "#password": "***"})
    page.click("button[type=submit]")
    page.wait_for("text=Dashboard")

    for cat in ["electronics", "furniture", "toys"]:
        page.navigate(f"https://competitor.example.com/prices/{cat}")
        prices = page.evaluate(
            "[...document.querySelectorAll('tr')].map(r => r.textContent)"
        )
        save_csv(cat, prices)

chrome.close()

Methods

| Method | Description | |---|---| | Chrome.connect() | Connect to or auto-start the SilbercueChrome server | | chrome.new_page() | Context manager — opens a new tab, auto-closes on exit | | page.navigate(url) | Navigate and wait for load | | page.click(selector) | Click element by CSS selector, text, or ref | | page.type(selector, text) | Type text into an input | | page.fill({"sel": "val"}) | Fill multiple form fields at once | | page.wait_for(condition) | Wait for JS condition or "text=..." shorthand | | page.evaluate(expression) | Run JavaScript, return result | | page.download() | Enable downloads, return download dir | | page.close() | Close the tab (auto-called by context manager) | | page.cdp.send(method, params) | Escape Hatch — direct CDP access via WebSocket (see below) |

Escape Hatch: Direct CDP Access

For use cases the high-level API doesn't cover — network interception, console log subscriptions, performance tracing, cookie management — you can drop down to raw CDP commands:

with chrome.new_page() as page:
    page.navigate("https://example.com")

    # Enable network tracking
    page.cdp.send("Network.enable")

    # Get all cookies
    cookies = page.cdp.send("Network.getAllCookies")

    # Performance tracing
    page.cdp.send("Tracing.start", {"categories": "-*,devtools.timeline"})

The Escape Hatch communicates directly with Chrome via WebSocket (port 9222), bypassing the server. It connects lazily on the first send() call and reuses the connection for subsequent calls. Each page gets its own WebSocket routed to the correct tab.

MCP Coexistence

When the MCP server and Python scripts need to run at the same time, add --script to the MCP config. Chrome.connect() handles the rest automatically — each script works in its own tab, MCP tabs are never touched.

Enabling `--script` in MCP Config

Claude Code:

claude mcp add --scope user silbercuechrome npx -y @silbercue/chrome@latest -- --script

Cursor / Cline (mcp.json):

{
  "mcpServers": {
    "silbercuechrome": {
      "command": "npx",
      "args": ["-y", "@silbercue/chrome@latest", "--", "--script"]
    }
  }
}

See python/README.md for the full API reference and advanced examples.

Free vs Pro

The Free tier gives you 18 tools that cover the entire benchmark suite. Pro adds virtual_desk, switch_tab, dom_snapshot, parallel tabs in run_plan, ambient-context hooks, and an operator hook pipeline on top — 23 tools in total.

| | Free | Pro | |---|---|---| | Tools | 18 | 23 | | Page understanding | view_page | view_page + dom_snapshot (spatial queries) | | Tab management | navigate, tab_status | + virtual_desk, switch_tab, parallel tabs in run_plan | | Interaction | click, type, fill_form, press_key, scroll, file_upload, handle_dialog | Same | | Observation | capture_image, wait_for, observe, console_logs, network_monitor | Same + ambient page-context hooks | | Scripting | run_plan (sequential) | run_plan (sequential + parallel + operator hooks) | | Last resort | evaluate | evaluate + anti-pattern scanner hints |

Pro costs €12/month. Get a license on Polar.sh, then follow Install Pro via Homebrew above — three commands, no manual download, no env-var editing. License keys arrive by email and are activated with silbercuechrome activate <YOUR-LICENSE-KEY>. (The SILBERCUECHROME_LICENSE_KEY=... env var still works as an alternative for non-Homebrew installs.)

Tool Overview

| Tool | Tier | Description | |---|---|---| | Reading & Observation | | | | view_page | Free | A11y-tree with stable e-refs — primary way to understand the page. Filter by interactive (default) or all. 5.4× more compact than Playwright's browser_snapshot. | | capture_image | Free | WebP screenshot, max 800px, <100KB. For visual verification only — refs come from view_page. | | console_logs | Free | Browser console output with level/pattern filters | | network_monitor | Free | Start/stop/query network requests with filtering | | observe | Free | Watch DOM changes: collect (buffer over time) or until (wait for condition, then auto-click) | | wait_for | Free | Wait for element visible, network idle, or JS expression true | | tab_status | Free | Active tab's cached URL/title/ready/errors (0ms) | | Interaction | | | | click | Free | Real CDP mouse events by ref, selector, text, or coordinates. Response includes DOM diff (NEW/REMOVED/CHANGED). | | type | Free | Type into an input by ref/selector | | fill_form | Free | Fill a complete form in one call — text, <select>, checkbox, radio. Per-field status. | | press_key | Free | Real CDP keyboard events — Enter, Escape, Tab, arrows, shortcuts (Ctrl+K, etc.) | | scroll | Free | Scroll page, element into view, or inside a specific container | | file_upload | Free | Upload file(s) to <input type="file"> | | handle_dialog | Free | Configure alert/confirm/prompt handling before triggering actions | | Navigation | | | | navigate | Free | Load a URL. First call per session auto-redirected to virtual_desk to prevent overwriting the user's tab. | | Scripting | | | | run_plan | Free (3) / Pro (∞) | Multi-step batch execution with variables, conditions, saveAs, error strategies, suspend/resume. Parallel tabs require Pro. | | configure_session | Free | View/set session defaults (tab, timeout) and accept auto-promote suggestions | | evaluate | Free | Execute JS in page context. Anti-pattern scanner warns on querySelector/.click(). | | Pro | | | | virtual_desk | Pro | Lists all tabs with stable IDs. Call first in every session. | | switch_tab | Pro | Open, switch to, or close tabs by ID from virtual_desk | | dom_snapshot | Pro | Bounding boxes, computed styles, paint order. For spatial questions view_page cannot answer. |

Benchmarks

Measured on https://mcp-test.second-truth.com — 35 tests in 5 levels (Basics, Intermediate, Advanced, Hardest, Community Pain Points). Each run is independent, values on the benchmark page are randomized per page-load, all runs started in a fresh Claude Code session out of /tmp (no project context bias), and all metrics measured post-hoc from the session JSONL via test-hardest/measure-tool-calls.sh — no self-reporting, no MCP-side instrumentation, just counting tool_use blocks and tool_result char lengths.

Head-to-Head (24-Test Suite, 2026-04-04)

All four servers ran the same 24-test suite on mcp-test.second-truth.com, same LLM (Claude Opus 4.6), same test page. Raw data in test-hardest/benchmark-*.json.

| MCP Server | Tests Passed | Duration | Tool Calls | Speed vs SC | |---|---:|---:|---:|---| | SilbercueChrome | 24/24 | 21s | 116 | -- | | Playwright MCP | 24/24 | 570s | 138 | 27x slower | | claude-in-chrome | 24/24 | 772s | 193 | 37x slower | | browser-use | 16/24 | 1813s | 124 | 86x slower |

Pass Rate + Duration (35-Test Suite, 2026-04-09)

| MCP | Passed | Duration | |---|---|---| | SilbercueChrome Free | 30/31 (97%) | 598s | | Playwright MCP | 29/31 (94%) | 563s | | Playwright CLI | 28/31 (90%) | 376s |

Pending re-bench on the new 35-test suite: SilbercueChrome Pro, chrome-browser, claude-in-chrome, browser-use, Browser MCP. The previous 24-test results are archived in the git history.

Tool-Efficiency (the fair metric)

We measure each tool call's response char length directly, group by tool name, estimate tokens via chars/4. Why this metric: session-level token deltas are dominated by LLM overhead (system prompt + CLAUDE.md + conversation history = ~80-90% of the budget) and only show 5-15% differences between MCPs — untrustworthy for comparing browser servers. Tool-response size is the part the MCP server actually controls.

| Metric | SC Free | Playwright MCP | Difference | |---|---:|---:|---:| | Tool calls (MCP-only) | 151 | 121 | +25% (SC uses more, smaller calls) | | ∅ Response size | 807 Chars | 1.448 Chars | SC 1.8× smaller | | ∅ Response tokens est. | 201 | 362 | SC 1.8× smaller | | P95 Response | 2.328 Chars | 8.068 Chars | SC 3.5× smaller | | Total response content | 128k Chars | 175k Chars | SC 27% less |

Per-Tool Breakdown (where the difference comes from)

| Tool | SC Free ∅ | Playwright MCP ∅ | Verdict | |---|---:|---:|---| | view_page / browser_snapshot | 1.124 Chars (21 calls) | 6.084 Chars (8 calls) | SC 5.4× more compact per call | | evaluate / browser_evaluate | 510 Chars (33 calls) | 2.155 Chars (47 calls) | SC 4.2× more compact per call | | type / browser_type | 88 Chars (13 calls) | 147 Chars (13 calls) | SC 1.7× more compact | | click / browser_click | 1.278 Chars (63 calls) | 463 Chars (44 calls) | Playwright 2.8× leaner — but see trade-off below |

The Ambient-Context trade-off

Ambient Context — Claude sees DOM changes for free, no extra view_page needed

SC's click is 2.8× larger than Playwright's because every click response embeds the DOM diff (NEW/REMOVED/CHANGED lines). Playwright returns a bare confirmation, forcing the LLM to follow up with a browser_snapshot or browser_evaluate to see what happened. Over a full benchmark run, this cascade costs Playwright MCP 47 extra browser_evaluate calls averaging 2.155 chars each. SC delivers the diff inline. Net result: SC's click+read_page+evaluate total is 120k chars vs Playwright MCP's 170k — 30% less response content overall.

view_page is 5.4× more compact than Playwright MCP's browser_snapshot

Measured on the 35-test benchmark (2026-04-09): SC's view_page averages 1.124 chars per call vs Playwright MCP's browser_snapshot at 6.084 chars. Same page, same test suite, same LLM driver. The a11y-tree compression + Ambient Context pipeline means we only send what the agent actually needs — smaller responses, less context pressure, cheaper runs.

See test-hardest/BENCHMARK-PROTOCOL.md for the full protocol, per-test breakdown, and raw JSON runs with tool_efficiency blocks.

Architecture

SilbercueChrome (Node.js MCP server, @silbercue/chrome)
├── @modelcontextprotocol/sdk (stdio transport)
├── CDP Client
│   ├── WebSocket transport (existing Chrome on :9222)
│   └── Pipe transport (auto-launched Chrome with --remote-debugging-pipe)
├── Auto-Launch: Chrome + optimal flags, visible by default
├── A11y-tree cache + Selector cache
├── Session Manager (OOPIF support for iframes and Shadow DOM)
├── Tab State Cache (URL/title/ready across tabs)
├── Script API (Python, pip install silbercuechrome)
│   ├── Shared Core via HTTP (:9223) — same tool handlers as MCP
│   └── Escape Hatch via WebSocket (:9222) — direct CDP for power users
└── 18 Free-tier tools + 3+ Pro-tier tools
    Reading · Interaction · Navigation · Scripting · Observation

Connection priority:

Auto-Launch (default, zero-config) — starts Chrome as a child process via --remote-debugging-pipe, visible as a window, with all flags set for reliable screenshots and keyboard focus.
WebSocket (optional) — if you already run Chrome with --remote-debugging-port=9222, SilbercueChrome connects to that instead. Use this to control your own browser with its extensions and login sessions.

Requirements

Node.js >= 18
Google Chrome, Chromium, or any Chromium-based browser (auto-detected on macOS/Linux/Windows; override with CHROME_PATH)

Environment Variables

| Variable | Values | Default | Description | |---|---|---|---| | SILBERCUE_CHROME_AUTO_LAUNCH | true / false | true | Auto-launch Chrome if no running instance found | | SILBERCUE_CHROME_HEADLESS | true / false | false | Opt-in headless mode for CI/server environments | | SILBERCUE_CHROME_PROFILE | path | — | Chrome user profile directory (auto-launch only) | | SILBERCUE_CHROME_FULL_TOOLS | true / false | false | Expose the full tool set (21 tools) instead of the leaner default set (10 tools) in tools/list | | CHROME_PATH | path | — | Path to Chrome binary (overrides auto-detection) | | SILBERCUECHROME_LICENSE_KEY | license key | — | Pro license key (e.g. SC-PRO-...) |

Known Issues

BUG-003: WebSocket `Sec-WebSocket-Accept` Mismatch (Node 22 + Chrome 146)

When connecting to an already-running Chrome via --remote-debugging-port=9222 (WebSocket transport), Node 22's undici 6.21.1 produces a different Sec-WebSocket-Accept hash than Chrome 146 expects. This is a confirmed bug in Node 22's native WebSocket implementation.

Workaround: The Accept validation is skipped — safe because the connection is to a localhost CDP endpoint. The workaround is already active in the shipped code.

Auto-Launch is NOT affected. The default mode (auto-launch) uses --remote-debugging-pipe which bypasses WebSocket entirely. You only hit this if you manually start Chrome with --remote-debugging-port and connect via --attach.

License

The core server and all 18 Free-tier tools are MIT licensed — see LICENSE. Use them however you want, commercially or otherwise.

Pro tools (3+ gated tools, parallel tab execution, ambient context, operator hooks, faster internals) require a paid license. The license validation code is in the separate private Pro repository.

Contributing

Issues and pull requests welcome at github.com/Silbercue/silbercuechrome.

Privacy

SilbercueChrome runs entirely on your machine. All browser automation happens locally via CDP. No telemetry, no remote calls, no data sent to any third party.