Urgent.News

What's breaking now, across thousands of outlets.

AI

kywai: an open-source workspace for Claude Code, Codex and Gemini

Switching between coding agents also means moving context between them, deciding who handles each task, and working out what changed afterward. I'm building kywai to bring that workflow into one desktop app. It is free and MIT-licensed, with downloads for Windows, macOS and Linux. I'm looking for developers who already use coding agents to try a small task and tell me where the workflow breaks…

Kywai is an open-source desktop app designed to streamline the workflow of using multiple coding agents like Claude Code, Codex CLI, and Gemini CLI. It is free and available for Windows, macOS, and Linux, with sources available on GitHub for developers to inspect.

Kywai brings together these agents into a shared project workspace, where a coordinator proposes a plan and delegates tasks to the agents. The app can also ask another agent to review the changes made. For agents that use CLIs, Kywai runs the CLI with the agent's own access.

The interface offers a chat, task progress, file changes, and even a live 3D world where each agent’s activity is represented by a robot. Users can directly interact with an agent if a task doesn’t require a coordinated plan.

One aspect that requires feedback is the process of reviewing changes. While the app has Git-backed task diffs and undo functionality, showing changes after they’ve been applied, the current system does not provide a pre-apply approval screen. The question is whether the combination of task progress, file diffs, and another agent's review is sufficient to keep working or if users still need to reconstruct the entire sequence themselves.

To test the workflow, the developer suggests trying a small task with a supported agent. Users are asked to ask one agent to make a small change, have another agent review the result, and inspect the changed files. Comparing this experience with the normal workflow is encouraged. Feedback is sought on the ease of setup, clarity about each agent's actions, and whether the same workflow would be used for other tasks.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at dev.to →

More in AI

Is Mandating AI Use in the Workplace Backfiring?

Some workplaces are holding mandatory AI training sessions, while others "are requiring employees to use a specific number of tokens," reports the Washington Post, "or log a certain amount of time on…

  • Mandating AI use may backfire, according to Stanford professor Arvind Karunakaran
  • Consuming many tokens doesn't guarantee substantial value creation
  • Employees may feel undervalued and disengaged without considering benefits

GitHub Copilot updates for JetBrains and Xcode

Copilot Is Everywhere Now. The Supervision Problem Is Still Yours. GitHub just pushed Copilot into JetBrains Air and Xcode's built-in coding assistant in the same week, and the pitch is the same one…

More from Saturday 7 November →