Agy Bridge

Name: Agy Bridge
Author: sshahzaiib

@sshahzaiib

About Agy Bridge

MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI (agy) — purpose-built tools, model routing with fallback, session continuity, and output truncation to save Claude's context and tokens.

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "agy-bridge": {
      "command": "npx",
      "args": [
        "-y",
        "agy-bridge"
      ]
    }
  }
}

Tools

Delegate file analysis to the Antigravity CLI (Gemini) instead of reading files yourself. USE THIS whenever a file is large (>200 lines) or the task spans more than 3 files: logs, database dumps, generated code, cross-file reviews, comparisons. The files never enter your context — only the answer does.

Delegate codebase archaeology to the Antigravity CLI: git log/diff/blame spelunking, wide greps across a repo, 'when/why did X change', 'where is Y used'. USE THIS instead of running many search commands yourself — it saves your context.

Delegate a web/documentation lookup to the Antigravity CLI (Gemini with web access): library docs, API references, error messages, current versions, external knowledge. USE THIS when you need information you don't have or that may be newer than your training data.

Get an adversarial second opinion from a different model family (Gemini Pro). ALWAYS use this for plan critiques, design reviews, and pre-merge code review: it hunts for flaws, edge cases, security issues, and unstated assumptions you may have missed.

Continue a previous Antigravity session by session_id (returned by every other tool). USE THIS for follow-up questions about a prior delegation — the full prior context is already on agy's side, so you don't resend anything.

Raw delegation to the Antigravity CLI for heavy tasks that don't fit the other tools. agy has full tool access (shell, file reads, web) in the given cwd.

Overview

What is Agy Bridge?

An MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI (agy). Claude sends a task, the bridge routes it to the best available model via agy, and only the answer returns — saving Claude's context window and tokens.

How to use Agy Bridge?

Register the MCP server with Claude Code using claude mcp add -s user agy-bridge npx -- -y agy-bridge, then add delegation rules by downloading the recommended CLAUDE.md file. The bridge exposes six tools; Claude self-routes tasks based on their type.

Key features of Agy Bridge

6 purpose-built tools for different task types
Per-tool model routing with availability detection and fallback
Session continuity via the follow_up tool
Configurable output truncation cap (default 50,000 characters)
Optional sandbox mode for safety
Zero-install npx usage (Node.js)

Use cases of Agy Bridge

Analyze large files, logs, or generated code without filling Claude’s context
Perform deep git archaeology (log, diff, blame) across repositories
Look up external documentation, API references, or current knowledge via web
Conduct adversarial code and design reviews using high-capability models
Delegate any heavy or complex task to a dedicated model via delegate

FAQ from Agy Bridge

What are the requirements for Agy Bridge?

Node.js 18+, the Antigravity CLI (agy) installed and authenticated, and Claude Code.

How does model routing work?

On first use the bridge runs agy models (cached per process) and picks the first available model in each tool’s preference chain. If none is available it falls back to AGY_DEFAULT_MODEL, then to agy’s own default.

How is Agy Bridge different from claude-to-agy?

Agy Bridge provides 6 purpose-built tools instead of 1 generic tool, per-tool model routing with fallbacks, session continuity via follow_up, configurable output truncation, and an optional sandbox mode.

Can I configure output truncation?

Yes, set the AGY_MAX_OUTPUT_CHARS environment variable (default 50,000) to cap the tool output and protect Claude’s context.

Is there session support for multi-turn conversations?

Yes, the follow_up tool accepts a session_id returned in the footer of every response, allowing you to continue prior conversations without resending context.

Comments

More Developer Tools MCP servers

Serena

oraios

A powerful MCP toolkit for coding, providing semantic retrieval and editing capabilities - the IDE for your agent

26.2K

mcp-excalidraw

yctimlin

MCP server and Claude Code skill for Excalidraw — programmatic canvas toolkit to create, edit, and export diagrams via AI agents with real-time canvas sync.

2.1K