Termexo

<p align="center"> <img src="apps/desktop-ui/public/termexo-mark.svg" width="104" alt="Termexo logo"> </p> <h1 align="center">Termexo</h1> <p align="center"><strong>One window for every coding agent</strong></p> <p align="center"> <strong>English</strong> · <a href="./README.cn.md">简体中文</a> </p> <p align="center"> <img alt="Version 0.7.0" src="https://img.shields.io/badge/version-0.7.0-58c7a0"> <img alt="Windows" src="https://img.shields.io/badge/platform-Windows-0078D4?logo=windows"> <img alt="Tauri 2" src="https://img.shields.io/badge/Tauri-2-24C8DB?logo=tauri&logoColor=white"> <img alt="Angular 22" src="https://img.shields.io/badge/Angular-22-DD0031?logo=angular"> </p> <p align="center"> <a href="https://www.termexo.com">Website</a> · <a href="https://github.com/gemron/Termexo/releases/latest">Download</a> · <a href="https://www.npmjs.com/package/termexo">npm</a> </p>

Termexo puts Claude Code, Codex, and the terminals around them into one recoverable Windows workspace. Keep several agents visible at once, see immediately when one needs input or approval, reopen yesterday's native session, and switch the Claude CLI between compatible model providers without rebuilding your setup.

Run the complete Windows app with one command—no Termexo account or server required:

npx termexo@latest

The current version is V0.7.0. The window drops the system title bar for chrome of its own, with the top bar spanning the width and both side panels beneath it; terminals render on the GPU so a long scrollback scrolls smoothly; a terminal keeps its account across reconnects and restarts; and configuration, plugins, and skills can be copied between accounts without moving any credentials.

<p align="center"> _{Claude Code and Codex terminals side by side in a workspace that remembers its layout.} </p>

What Works Today

<table> <tr> <td width="50%" valign="top"> <strong>Four agents. One screen.</strong><br><br> Run Claude Code, Codex, and OpenCode side by side in as many real PTY terminals as you need, choose which stay visible, and arrange them in a custom 1–6 row/column grid. Reorder tabs by dragging, close one with the middle mouse button, and drive the workbench from the keyboard. Each workspace remembers its folder, tabs, layout, model, and theme. <br><br> <a href="website/assets/termexo-workbench.png"><img src="website/assets/termexo-workbench.png" alt="Termexo multi-agent workbench"></a> </td> <td width="50%" valign="top"> <strong>Know when an agent needs you.</strong><br><br> Waiting for input, waiting for approval, completed, and failed states are distinct. A persistent banner, native Windows notification, and taskbar flash bring you back to the exact terminal that needs attention. <br><br> <a href="website/assets/termexo-attention.png"><img src="website/assets/termexo-attention.png" alt="Termexo agent attention notifications"></a> </td> </tr> <tr> <td width="50%" valign="top"> <strong>Pick up yesterday's conversation.</strong><br><br> Search local Claude Code, Codex, and OpenCode sessions across projects, accounts, branches, and models. Termexo restores them through the CLIs' own <code>claude --resume</code>, <code>codex resume</code>, and <code>opencode --session</code> commands, reclaims a Claude session the CLI still holds open, and keeps native session files read-only. <br><br> <a href="website/assets/termexo-session-center.png"><img src="website/assets/termexo-session-center.png" alt="Termexo native session center"></a> </td> <td width="50%" valign="top"> <strong>Same CLI. Different model.</strong><br><br> Point Claude Code at Anthropic, DeepSeek, MiniMax, GLM, or a custom Anthropic-compatible endpoint. Save providers as profiles, keep API keys in Windows Credential Manager, and switch every Claude terminal in a workspace together. <br><br> <a href="website/assets/termexo-models.png"><img src="website/assets/termexo-models.png" alt="Termexo model provider profiles"></a> </td> </tr> <tr> <td width="50%" valign="top"> <strong>Turn a task into a running agent.</strong><br><br> The task board keeps projects and tasks with priorities and acceptance criteria. Hand one to Claude Code, Codex, or OpenCode and it becomes a real terminal, moving through todo, executing, completed, and verified as that terminal reports its own status. </td> <td width="50%" valign="top"> <strong>Let it run without babysitting.</strong><br><br> Start any agent with automatic confirmation — <code>--permission-mode auto</code> for Claude, <code>--approve-for-me</code> for Codex, <code>--auto</code> for OpenCode — so the AUTO chip on a terminal means the same thing whichever agent drew it. </td> </tr> </table>

Termexo is local by default: it has no account, cloud service, or required sync. Workspace state stays in local SQLite storage, credentials stay in Windows Credential Manager, and Claude/Codex session histories are read-only. The agent CLIs still connect to the providers you configure under their own terms and privacy policies.

The interface is available in Simplified Chinese, English, Spanish, French, German, Japanese, and Korean. It follows the Windows language automatically, or you can choose a language from the main toolbar and keep that choice across restarts.

What's New in V0.7.0

• The window draws its own chrome. No system title bar: the top bar spans the whole window with the window controls at its right edge, and both side panels start beneath it. Dragging the bar still moves the window, double-clicking still maximises.
• Terminals render on the GPU. A long scrollback scrolls without the stutter the DOM renderer produced. Machines without a usable GPU fall back to the previous renderer.
• A terminal keeps its account. Reconnecting — including after the app restarts — rebuilds the account directory, proxy settings, and provider key from what the terminal records, instead of quietly falling back to the CLI's default home.
• Switch a terminal's account from its header. The header names the account it runs on, and picking another restarts that terminal on it with a new session, leaving the model, MCP profile, and automatic confirmation alone.
• Copy configuration between accounts. Settings, instructions, plugins, and skills move across; credentials, account identity, and session history never do.
• Sign-in is noticed on its own. A finished login refreshes the account without waiting for a CLI that keeps running after the browser flow returns.

<p align="center"> _{A task carries its acceptance criteria from todo through to verified, and runs as a real agent terminal.} </p>

Release notes for every earlier version live in CHANGELOG.md.

Why Termexo

AI coding tools usually run in isolated terminals and sessions. As the number of projects grows, developers have to remember:

• which terminal belongs to which project, branch, and agent;
• which sessions are running, waiting for input, or waiting for approval;
• how to resume a specific Claude session;
• how model, endpoint, API key, and MCP configurations fit together;
• which parts of a workspace can actually be restored after an application restart.

Termexo uses the Workspace as its organizing unit and turns this scattered state into a local control plane that is observable, recoverable, and extensible.

Implemented Features

| Capability | Current implementation | | ------------------------ | ------------------------------------------------------------------------------------------------------------------ | | Workspace management | Create, rename, theme, manually reorder, and switch workspaces; persist paths, layouts, and terminal configuration | | Multi-terminal workbench | Unlimited tabs, explicit pane selection, configurable 1–6 row/column grids, pane/workspace maximize, and real PTYs | | Agent detection | Detect the Claude Code, Codex, and OpenCode executables, their versions, and health on Windows | | Start agent sessions | Launch Claude, Codex, or OpenCode with a working directory, isolated login account, Agent-specific model configuration, and optional automatic confirmation | | Session center | Read-only multi-account Claude/Codex/OpenCode discovery, search, workspace filtering, native resume, and reclaim of a Claude session the CLI still holds open | | Agent status tracking | Isolated hooks per terminal for thinking, tool use, approval, user input, completion, and failure states | | Model and MCP profiles | Manage endpoints, keys, and MCP configuration; switch Claude CLI across Anthropic-compatible backends | | Network and npm profiles | Scope HTTP/HTTPS/SOCKS and npm settings globally or per workspace, test reachability, and inject them at launch | | Account management | Manage multiple isolated Claude and ChatGPT/Codex logins, defaults, authentication status, and launch-time choice | | Managed CLI lifecycle | Preview, confirm, install, or upgrade the official Claude Code, Codex, and OpenCode npm packages, then verify the result | | Task board | Organise projects and tasks with priorities and acceptance criteria, run one as a Claude/Codex/OpenCode terminal, and track it from todo through executing, completed, and verified | | Prompt assets | Recover live per-terminal drafts; search, favorite, pin, delete, and reuse submitted prompts | | Session handoff | Build redacted, token-budgeted Git/task packages; import/export documents and continue in another Agent | | Local data and secrets | Store workspace/session/event data in SQLite and API keys in Windows Credential Manager | | Browser preview | Preview the complete UI without Rust and exercise layout flows through an interactive simulated terminal |

<p align="center"> _{Models, endpoints, and credential entry are managed in one place. Stored secrets are never returned to the frontend.} </p>

Design Goals

1. Local first — Project paths, terminals, session indexes, and configuration stay on the local machine by default. Termexo does not require a Termexo cloud service.
2. Preserve native agent semantics — Prefer native session and configuration mechanisms instead of pretending that a terminated process is still alive.
3. Coordinate terminals, do not replace them — Termexo provides the workbench, state, and orchestration layer. Commands still run inside real PTYs and agents.
4. Make agent state observable — Map agent-specific events into common running, thinking, input, approval, completed, and failed states.
5. Keep security boundaries explicit — Secrets belong in the operating-system credential store, not in SQLite, snapshots, hook payloads, or logs.
6. Grow into a multi-agent architecture — Evolve around adapters, PTY, hooks, snapshots, and routing boundaries so more CLIs can be added without rebuilding the core.
7. Extend workspaces across trusted devices — Add sharing, remote access, mobile approvals, and collaboration without weakening local ownership or security boundaries.

Current Boundaries

• Claude Code and Codex both support native detection, account/mo