browser-automationplaywrightmcphuman-in-the-loopaudit
Runs Playwright browser sessions over MCP with noVNC takeover, reusable auth profiles, approvals, and audit records.
Quick Install
Automated & IDE SetupCopy the AI prompt to install this server into Claude Code, Cursor, or another agent — or use 1-click editor setup below.
Not yet automatically verified
We haven't yet run this listing's install command through our automated sandbox check. This isn't a red flag — we're steadily working through the catalog.
Manual Client & Custom JSON ConfigExpand JSON ▾
Client Config & Setup
Choose your client or environment
Target File:
💡 Paste the JSON block into your client's configuration file under mcpServers, then restart the application.
Overview
Use cases
Automate internal dashboards and admin tools
Recover brittle browser flows with human takeover
Reuse authenticated sessions across browser tasks
Inspect pages, downloads, and network activity
Audit agent-driven browser operations
Key features
Playwright browser sessions over MCP
noVNC human takeover
Reusable named authentication profiles
Approval and audit controls
HTTP, stdio, and REST access
Docker Compose local deployment
Capabilities & Tool Schemas
Inspect callable tools, capabilities, and parameters exposed to AI agents by Auto Browser.
Extracted Tool Capabilities
Playwright browser sessions over MCP
noVNC human takeover
Reusable named authentication profiles
Approval and audit controls
HTTP, stdio, and REST access
Docker Compose local deployment
How Auto Browser works
What LvcidPsyche/auto-browser MCP server does
LvcidPsyche/auto-browser MCP server provides a browser control plane for authorized workflows. Agents can create and observe Playwright-backed sessions, interact with pages, manage tabs, inspect network activity, capture downloads, and work with screenshots, DOM summaries, OCR excerpts, or extracted text. The system also exposes operator-facing controls so a person can view and continue the same session through noVNC.
A central use case is login-once, reuse-later automation. An operator can open a session, complete a login manually when required, save the resulting state as a named authentication profile, and start later sessions from that profile. This is suited to internal dashboards, admin tools, browser debugging, and sites where an automated flow may need human recovery.
How it works
The application is built around a Playwright browser and a FastAPI controller. Docker Compose starts the local stack, including the controller and browser environment. MCP clients can connect through the HTTP endpoint at http://127.0.0.1:8000/mcp, use the convenience tool endpoints, or run the bundled stdio bridge with uvx auto-browser-mcp.
LvcidPsyche/auto-browser MCP server offers a curated MCP tool profile by default. Deployments can select the full internal tool surface with MCP_TOOL_PROFILE=full. The repository also documents direct REST use, including session creation and observation, so curl-based callers do not need an MCP client.
The operator dashboard is available at port 8000, and the noVNC interface is available at port 6080 in the default local setup. Published ports bind to 127.0.0.1 by default. Codespaces can provision the stack automatically for a hosted development environment.
Setup and configuration
Clone the repository, change into its directory, and start the stack with:
The README identifies this as sufficient for local development with default settings. An optional .env file can be created from .env.example, and make doctor checks the local setup when run from a terminal with Docker access and permission to open localhost sockets.
For the MCP stdio path, install or run the published bridge with:
The controller still needs to be available for the bridge to connect to. Claude Desktop, Cursor, other HTTP or stdio-capable MCP clients, and direct REST callers are listed as supported connection styles.
Tools and capabilities
LvcidPsyche/auto-browser MCP server includes documented surfaces for:
- Creating browser sessions and opening a starting URL.
- Observing pages with accessibility information, extracted text, and interactables.
- Finding page content with
browser.find_elements, using plain text or regular-expression queries with case-insensitive matching and surrounding context. - Controlling tabs and using screenshots, DOM summaries, OCR excerpts, downloads, and network inspection.
- Saving and reusing named authentication profiles.
- Reviewing recent audit events through the
browser://audit/eventsMCP resource. - Applying approvals, operator identity headers, PII scrubbing, protection profiles, and signed Witness receipts.
- Using noVNC for human takeover of a live browser session.
The repository also describes staged skill candidates derived from verified browser traces, with provenance checks and review-only graduation when mesh identity is configured.
Limitations and notes
The project is intended for authorized browser workflows. Its stated non-goals include CAPTCHA solving, unauthorized scraping or account automation, and deceptive identity shaping or bypass tooling. Human takeover is available for brittle flows, but it does not turn those unsupported activities into supported use cases.
The default text observation preset returns accessibility structure, extracted text, and interactables without screenshots or OCR. Deployments can set PERCEPTION_PRESET_DEFAULT=text to make that preset the default. The local quickstart exposes API documentation at /docs, a dashboard at /dashboard, and noVNC at /vnc.html?autoconnect=true&resize=scale on the documented localhost ports.
Related MCP Servers
Adoption & maintenance
Factual signals from GitHub, npm, and our automated checks — not a rating.
GitHub stars
785
Stargazers on the source repository.
Last commit
11d ago
Most recent push to the default branch.
Directory activity
3 views
Config copies, upvotes, and views on AllMCPs.
Reviews
No reviews yet — be the first to share how this listing worked for you.
Frequently Asked Questions about Auto Browser
AllMCPs Directory Badge
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
Badge Style:
Live Dynamic SVG Preview
Markdown (GitHub README)
[](https://allmcps.com/mcp/lvcidpsyche-auto-browser)HTML Embed
<a href="https://allmcps.com/mcp/lvcidpsyche-auto-browser"><img src="https://allmcps.com/api/badge/lvcidpsyche-auto-browser?style=directory" alt="Auto Browser on AllMCPs" /></a>Technical Specs & Signals
CategoryBrowser Automation
PricingFree
10/10 checks healthy over the last 30d
59Quality signal: Good · 59/100How this signal is calculated ▾
Server availabilityNot measured
Not scored for repo-hosted servers — we can't reach the running server, only its GitHub page. Hosted MCP endpoints are health-checked live.
Verified ownership10/20
Documentation & tools24/30
Adoption & activity10/15
Community engagement0/10
A guidance signal from public completeness & health data — not a user rating. New listings start lower and rise as they add docs, get verified, and grow adoption. Signals we can't observe for a listing are skipped, not counted against it.
Supply-chain signal
No high-severity advisories surfaced by our automated scan.
Critical 0High 0Medium 0Low 0
Scanned 20d ago via OSV.dev · auto-browser-mcp (PyPI)
Own this project?
This directory is pre-filled from public sources. Claim via GitHub README, site badge, or DNS TXT to unlock edit access and the Official badge — proof is checked automatically, then reviewed by our team.
Free dofollow backlink: add your website and place the AllMCPs badge on it — no claim needed. We detect it automatically and keep it verified as long as the badge stays live.
Claim & get free dofollowShare & Embed
Add our SVG badge (dark/light directory styles) or embeddable widget to your site.