Existing workbench screenshot (v0.8.0), shown for context.
I'm the maintainer of Termexo, an MIT-licensed Windows workspace for Claude Code, Codex and OpenCode.
v0.8.7 focuses on a frustrating reliability problem: an agent could keep running while its terminal went blank and appeared stopped. Resizing around a wide character could trigger a parser panic, poison the screen mutex and leave subsequent screen reads failing. Switching between clients could reproduce that combination.
I maintain Termexo, an MIT-licensed Windows workbench for Claude Code, Codex and OpenCode. Version 0.8.6 focuses on reloading a view, reconnecting remotely, resizing a window and switching to a terminal that was hidden.
Keep the running process Previously, frontend loading could be mistaken for a new terminal launch. Reloading could restart an agent that was already working. The view now reconnects to a backend terminal that is still alive, preserving its current launch and avoiding redundant session takeover requests. Closing Termexo or shutting down the computer still ends running processes.
Maker update: Termexo v0.8.5 is out. I maintain this MIT-licensed Windows workbench for Claude Code, Codex CLI, OpenCode and real terminals.
More consistent terminal behavior
The EXE, MSI and npm packages now bundle Microsoft's MIT-licensed ConPTY components. Older Windows ConPTY builds can lose Shift+Tab or mouse reports, affecting Claude Code shortcuts and OpenCode scrolling. Using a bundled version reduces these OS-version differences and adds about 600 KB to the installer.
The frontend also detects its pseudo console and avoids reflowing wrapped lines a second time when an older system ConPTY has already done it.
Maker update: Termexo v0.8.2 is out today. This release focuses on typing responsiveness while multiple agents produce output.
Output now uses one application-wide subscription with routing by terminal ID, reducing duplicate callbacks and UI work. History replay restores the terminal display without reprocessing old output through task, handoff, and startup readers. Draft persistence is batched, and terminal size ownership is handled on focus/click instead of every keystroke.
Also new: Claude 1M context configuration, separate Claude/Codex effort settings, per-terminal launch overrides, and direct agent buttons in empty workspaces. Actual context and effort support depends on the CLI, model, account, and provider.
Termexo is MIT-licensed and runs on Windows. Agent CLIs and model services need their own setup. Release notes and installers: https://github.com/gemron/Termex...
Maker update: Termexo v0.8.3 is out. I maintain this MIT-licensed Windows workbench for Claude Code, Codex CLI, OpenCode and real terminals. If typing has felt delayed while agents are working, this is an update worth installing.
Termexo brings Claude Code and Codex into one recoverable Windows workspace. Arrange real PTY terminals in custom grids, search and resume native sessions, get notified when an agent needs approval, and switch Claude-compatible model profiles without rebuilding environment variables. Local-first, no Termexo account required.
Run three coding agents in real PTY terminals, turn tasks into agent sessions, resume native history, hand work across agents, and get notified only when a terminal needs attention. Local-first and MIT licensed.
