Antigravity Argon Bridge
Allows querying Google's 4 Argon reasoning engine through linked Google Pro / Gemini Advanced accounts, with configurable thinking depth, multi-account rate-limit failover, and usage/quota inspection.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Antigravity Argon BridgeAsk Argon to review my MCP server design with high thinking depth"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Antigravity Argon Bridge [DEPRECATED / ARCHIVED]
Project Status & Deprecation Notice
As of October 26, 2026, Antigravity Argon Bridge has been permanently unpublished from the VS Code Marketplace and the Open VSX Registry, and this repository has been archived.
Why was this project taken down?
Antigravity Argon Bridge was created as an experimental bridge and proof-of-concept to provide developers and autonomous coding agents access to high-depth reasoning models through persistent browser sessions and CDP-driven automation.
While the project gained significant interest and adoption in a short period of time, automated browser driving and session emulation across Google accounts carry inherent security and account suspension risks for end users.
As maintainers, prioritizing the safety, privacy, and account security of our users outweighs popularity metrics. Rather than putting users at risk of sudden account blocks, session invalidations, or TOS flags, the decision was made to safely sunset the project and pull it from all extension registries.
Related MCP server: agy-subagent-mcp
A Huge Thank You to the Community
This was my very first open-source project, and the response from the developer and AI agent community was humbling:
550+ downloads on Open VSX
100+ downloads on the VS Code Marketplace
40+ GitHub stars
Valuable community bug reports and PR contributions
Thank you to everyone who tried it out, starred the project, gave feedback, and contributed.
Alternatives & Recommendations
If you are looking to integrate high-capability reasoning models into your development workflow:
Use official developer APIs (e.g. Google AI Studio / Gemini API) with standard API keys.
Leverage official MCP connectors that adhere strictly to provider-sanctioned API access.
This repository remains archived for historical purposes under the MIT License.
This server cannot be deployed
Maintenance
Related MCP Connectors
Cross-agent artifact workspace with provenance across Claude Code, Codex, Cursor, LangGraph.
Gives your AI assistant persistent memory and intelligence about your work patterns.
Agentic AI runtime: persistent memory, vault, autonomous agents, deep research, DeFi execution.
Create and manage AI agents that collaborate and solve problems through natural language interacti…
Related MCP Servers
- AlicenseAqualityAmaintenanceEnables Claude Code and Claude Desktop to delegate token-heavy tasks to Antigravity headless subagents, offloading file edits, test runs, and exploration while preserving Claude's context window.1212 npm1MIT
- AlicenseNot gradedqualityCmaintenanceEnables Claude Code to delegate tasks to Google Antigravity (Gemini) as an autonomous subagent, including codebase exploration, architecture analysis, multi-file code generation, and technical Q&A, without requiring external API keys.MIT
- AlicenseNot gradedqualityAmaintenanceEnables Codex users to delegate engineering, refactoring, auditing, and test execution tasks to Google Antigravity CLI as a persistent autonomous subagent, with tools for managing sessions and running prompts.1MIT
- AlicenseBqualityBmaintenanceEnables an AI to delegate work to secondary Antigravity CLI accounts, running jobs in parallel with dependency (DAG) scheduling, file read/write declarations, warm multi-turn sessions, account quota tracking and compact result collection. Users simply describe their request in natural language while the AI plans and the server executes — no hand-written workflow JSON required.21MIT