github.devotd.
NewCoding agents

nimbalyst/nimbalyst v0.81.7 — Visual workspace for parallel coding agents

nimbalyst/nimbalyst is an MIT-licensed open-source visual workspace that runs Claude Code, Codex, and OpenCode in parallel and lets you visually edit the resulting work as markdown, mockups, and diagrams.

Repositorynimbalyst/nimbalyst ↗

nimbalyst/nimbalyst is an MIT-licensed open-source visual workspace that runs Claude Code, Codex, and OpenCode in parallel and lets you visually edit the resulting work as markdown, mockups, and diagrams.

GitHub description: Nimbalyst - The open-source visual workspace for Claude Code, Codex, and OpenCode. Run multiple coding agents in parallel, edit their work visually in markdown, mockups, and diagrams, and track tasks. Free, MIT-licensed desktop app for macOS, Windows, Linux, with mobile companion for iOS and Android.

stars
★ 1.9K
Forks
312
Language
TypeScript
License
MIT
Latest release
v0.81.7
Read original source ↗View on GitHub ↗

nimbalyst/nimbalyst is an open-source visual workspace that runs several coding agents (AI tools that write code to perform tasks) such as Claude Code, Codex, and OpenCode at the same time, and lets you edit their results visually in one place. The latest release recorded is v0.81.7, distributed as a desktop app.

The README explains that it brings documents, code, tasks, and sessions together in one place and runs multiple agent sessions in parallel. Changes are reviewed as diffs (before-and-after comparisons), and multiple tasks are managed with worktrees and task boards. It is written in TypeScript and carries a free MIT license. It is a desktop app for macOS, Windows, and Linux, and the repository description states that companion apps for iOS and Android are provided as well. It has 1,945 stars.

However, no changelog or release notes for v0.81.7 appear in the record, so what this release changed cannot be known. The repository was created on October 30, 2025, and the only figures in the record are a momentum score of 4.3 (out of 10). No expressiveness, technical, or quality scores are given. Judging only by the scope described in the README, it reads as a tool that keeps several coding agents under control on a single screen, while actual performance and stability cannot be verified.