This project addresses the challenge of losing context when switching between AI coding assistants like Claude, Copilot, and Gemini. It provides a command-line tool, continues, that enables users to seamlessly pick up where they left off in one AI tool by transferring their session context (conversation history, file changes, working directory) to another. This is particularly beneficial for developers who encounter rate limits or wish to leverage different AI models, saving them the effort of re-explaining their project state.

How It Works

continues operates by discovering session directories for supported AI tools, parsing their native storage formats (JSONL, JSON, SQLite, YAML), and extracting crucial data. This includes messages, file edits, tool activity (like bash commands or API calls), and AI reasoning notes. The extracted information is summarized into a structured context document, which is then injected as a prompt into the target AI tool. This approach ensures a rich and accurate handoff, preserving the nuances of the previous session.

Quick Start & Requirements

• Installation: No installation required; run directly via npx continues. Alternatively, install globally with npm install -g continues.
• Prerequisites: Node.js 22+ is required.
• Supported Tools: Claude Code, GitHub Copilot, Gemini CLI, Codex, OpenCode, Factory Droid, and Cursor CLI.
• Documentation: Usage examples and command references are provided within the README.

Highlighted Details

• Cross-Tool Handoff: Supports 42 tested conversion paths between 7 AI coding tools, enabling seamless transitions.
• Auto-Discovery & Interactive Picker: Automatically scans session directories and provides an interactive TUI for browsing, filtering, and selecting sessions.
• Comprehensive Context Extraction: Captures detailed tool activity, AI reasoning, model information, and token usage from sessions.
• Scriptable Output: Offers JSON/JSONL output for integration into automated workflows and non-interactive use.

Maintenance & Community

The project is licensed under the MIT license and is attributed to Yigit Konur. No specific community channels (e.g., Discord, Slack) or detailed maintenance information beyond the development setup are provided in the README.

Licensing & Compatibility

• License: MIT.
• Compatibility: The MIT license generally permits commercial use and integration with closed-source projects.

Limitations & Caveats

The tool requires Node.js 22+ and the presence of the respective AI coding tool CLIs on the system. It relies on reading session data from the tools' native storage locations, and no explicit mention of alpha/beta status or known bugs is made.