Wayland puts 18 AI coding CLIs under one desktop app
The open-source project ties 18 agent CLIs, 178 workflows and a Rust engine into a command center that runs on your own machine.
In short
Wayland is a local-first desktop app that orchestrates 18 agent CLIs — among them Claude Code, Codex and Ollama — from a single surface.
At a glance
- Project page at the time of reading: 604 stars, 114 forks, 3 watchers and 133 open issues.
- Seven built-in tools: Read, Write, Edit, Bash, Grep, Glob and Spawn.
- Two licenses: the desktop app under GNU AGPL-3.0, the Rust engine Fuigo under Apache-2.0.
- Memory is split across five SQLite partitions: working, episodic, semantic, procedural and a model of the user.
- Stated requirements: Bun 1.3.x and Node 22 to 24, with installers for macOS, Windows and Linux.
Wayland is a local-first desktop app that orchestrates 18 agent CLIs from a single surface. FerroxLabs pitches it as a command center where planning, building and shipping happen in one place instead of across half a dozen terminals. At the time we read the page it showed 604 stars, 114 forks, 3 watchers and 133 open issues.
The loop the name promises
Perceive, reason, act, evolve — the repository builds its pitch around those four verbs. Perceiving means reading the request and the state of the machine; reasoning means picking a model for the job at hand. Acting runs through seven built-in tools: Read, Write, Edit, Bash, Grep, Glob and Spawn. The fourth step is an eval harness meant to sharpen skill prompts and keep the improvement.
How well that last step performs is not something the page answers. It publishes no benchmark figures and no measurement runs, so the claim currently rests on the project's own description.
What gets wired together
The supported set covers 18 ACP CLI agents, including Claude Code, Codex, Qwen Code, Goose and local Ollama instances, with Gemini connected natively. Everything runs on the user's own API keys rather than a hosted account. An optional Flux Router handles routing across models, and 178 ready-made workflows ship with the app. Model Context Protocol integrations extend it further.
Memory, rules, voice
Project memory is split across five SQLite partitions: working, episodic, semantic, procedural and a model of the user. A rulebook the project calls a constitution is applied across every connected agent. Voice input and output, plus specialist teams that assemble themselves, round out the feature list.
Licensing and lineage
The desktop app carries the GNU AGPL-3.0; the Rust engine, Fuigo, is Apache-2.0. Fuigo is a fork of xai-org/grok-build, and parts of the app descend from AionUi, with upstream attribution kept in the repository's notices.
Running it calls for Bun 1.3.x and Node 22 to 24. Installers are offered for macOS on both Apple Silicon and Intel, for Windows on x64 and ARM64, and for Linux as deb, rpm and AppImage.
What we could not confirm
An npm release is listed at version 0.12.4, but the package name could not be pinned down with confidence, so we leave it out rather than guess. No independent reporting on Wayland exists yet. Every figure above therefore traces back to a single source: the project's own repository page.
FAQ
What is Wayland by FerroxLabs?
A local-first desktop app that brings 18 agent CLIs, 178 ready-made workflows and a Rust engine together under one interface.
Which AI CLIs does Wayland support?
It lists 18 ACP CLI agents, among them Claude Code, Codex, Qwen Code, Goose and local Ollama instances, with Gemini connected natively.
What license is Wayland released under?
The desktop app uses the GNU AGPL-3.0 and the Fuigo engine uses Apache-2.0. Fuigo is a fork of xai-org/grok-build.
