Quil: Reboot-Proof Terminal Multiplexer with Persistent Claude Code Sessions
Key Takeaways
- ▸Survives host reboots with workspace restoration in under 30 seconds—a gap that tmux and Zellij don't fill
- ▸Claude Code sessions auto-resume with full context preserved; no manual context rebuilding or copy-paste required
- ▸Type-aware pane system handles five built-in types with hot-reloadable custom TOML plugins for additional tools
Summary
Quil is a terminal multiplexer designed for developers running complex multi-tool workflows alongside AI coding assistants. Unlike traditional multiplexers like tmux and Zellij, which survive network disconnects but not host reboots, Quil persists an entire workspace—including tabs, panes, layouts, working directories, and AI session IDs—across host reboots, snapping everything back in under 30 seconds. The tool features first-class Claude Code support with auto-resume functionality that preserves conversation context and working directory, eliminating the need to re-explain work or rebuild SSH tunnels. Quil also ships an MCP (Model Context Protocol) server that exposes 18 tools, allowing any MCP-capable client (Claude Desktop, Cursor, Claude Code) to read pane output, send keystrokes, and query workspace state. The tool introduces type-aware panes that understand how different tools (Claude Code, SSH, terminal, Stripe, OpenCode) resume differently, with custom TOML plugins supporting other AI tools.
- MCP server integration makes the terminal addressable by AI agents with 18 native tools for workspace control
- Cross-platform support including Windows clipboard workaround for Claude Code image pasting
Editorial Opinion
Quil fills a genuine gap at the intersection of terminal multiplexing and AI-assisted development. While competitors like herdr and Agent of Empires pursue broader agent orchestration with web dashboards, Quil's bet on native clients, MCP integration, and terminal-native persistence reflects a philosophy that AI tools should enhance existing developer workflows rather than replace them. As Claude Code and similar tools become central to development practices, persistence mechanisms like auto-resume and workspace snapshots may become table-stakes for any tool ecosystem claiming to support AI-driven work.
