Skip to main content

AionUi Developer Guide

What is this about?

AionUi is a local open-source cowork desktop where multiple AI assistants can work side by side on coding, documents, spreadsheets, slides, and automations. This guide explains how AionUi is positioned, how setup works, how the desktop runtime is shaped, and where to extend it.

Checked against primary sources

This guide is based on the official AionUi website and the official iOfficeAI/AionUi repository, checked on June 26, 2026.

In this section​

1. What AionUi is​

The shortest accurate description is:

AionUi is a local desktop workspace where multiple AI agents and assistant surfaces can actually cowork instead of living in separate apps and terminals.

The official site explicitly positions it around:

  • Claude Code,
  • Codex,
  • Gemini CLI,
  • 20+ built-in assistants,
  • direct model-provider access,
  • office and productivity work beyond coding.

2. The mental model​

Think of AionUi as five cooperating layers:

LayerWhat it does
desktop shellMain user-facing workspace
agent integrationsReuses or connects multiple agent engines
built-in assistantsProvide ready-made roles for office and creative work
automation and remote controlEnable scheduled and mobile-triggered workflows
provider and MCP wiringConnect models and tools without a required cloud relay

3. Why developers would choose it​

  • It is local-first and explicitly says it does not proxy chats through vendor servers.
  • It supports multiple agent ecosystems side by side.
  • It is broader than a coding assistant because it also targets slides, spreadsheets, documents, browser tasks, and 24/7 automation.
  • It tries to bridge the gap between developer agents and business-friendly desktop workflows.

4. What makes it unusual​

AionUi is not only a wrapper around one model. The official site presents it as:

  • one desktop,
  • multiple agent backends,
  • many built-in assistants,
  • remote control from messaging surfaces,
  • and cron-style background work.

That makes it feel closer to a local AI operations desk than to a single chat app.

  1. this overview,
  2. setup and first run,
  3. architecture and runtime,
  4. configuration and security,
  5. extending and contributing.

Sources​