Bridge

Name: Bridge
Author: Balachadal

0•

Facilitates shared conversation threads between Claude and Codex CLI models, leveraging local SQLite for context without requiring API keys.

When working with multiple AI coding assistants like Claude and Codex, maintaining context across different models can be challenging, leading to fragmented information. This tool addresses that by providing a multi-model conversation bridge that enables shared memory through local SQLite-backed conversation threads. It allows models to access the same history, fostering collaborative workflows where one model can analyze and another can implement, all without needing API keys, external orchestration, or cloud state. The server maintains these threads in a local SQLite database, allowing any MCP client to create, manage, and engage AI models with full, securely wrapped conversation context.

主な機能

01Shared conversation threads between Claude CLI and Codex CLI, stored in local SQLite

02Security-first design with anti-injection markers and `execFile`-only CLI calls

03Built-in self-test for validating setup and live model invocation

04Zero configuration, compatible with any MCP client

050 GitHub stars

068 MCP tools for thread management, context export, and model invocation

ユースケース

01Facilitating collaborative problem-solving between different AI models (e.g., analysis by Claude, implementation by Codex).

02Maintaining continuous context for debugging sessions across multiple AI interactions.

03Streamlining code review processes by allowing different AI models to access and contribute to a shared history.