AionUi - Free Open-Source Desktop Office Agent
AionUi Tutorial — A free open-source desktop office agent supporting various AI agents like Gemini CLI, Claude Code, Codex, and more. Integrates with 88API to enable multi-model workflows.
🚀 AionUi is a free, local, open-source Cowork, supporting various AI agents such as Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie. It provides a complete GUI interface and WebUI remote access functionality, serving as an open-source alternative to Cowork.
- Official Website: https://www.aionui.com
- GitHub Repository: https://github.com/iOfficeAI/AionUi
- Download Link: https://github.com/iOfficeAI/AionUi/releases

Core Features
💬 Multi-Session Chat
- Multi-session + Independent Context - Allows opening multiple chat sessions simultaneously, each with its own independent context memory.
- Local Storage - All conversation data is stored locally in an SQLite database and will not be lost.
🤖 Multi-Model Support
- Multi-Platform Support - Supports mainstream models like Gemini, OpenAI, Claude, Qwen, with flexible switching.
- Local Model Support - Supports local model deployment via Ollama, LM Studio, etc.
🤝 Multi-Agent Mode
- Run Multiple AI Agents Simultaneously - Allows running multiple AI agents concurrently (e.g., Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, etc.).
- Unified MCP Management - Unified management and configuration of all agents via Model Context Protocol (MCP), simplifying operation.
- Skills Configuration - Supports configuring exclusive Skills for different agents to extend their capabilities.
- Assistant Customization - Supports custom assistant configurations to create personalized AI workflows.
- Independent Configuration - Each agent can be configured and used independently without interference.
- Flexible Switching - Seamlessly switch between different agents to meet various scenario requirements.
🗂️ File Management
- File Tree Browsing + Drag-and-Drop Upload - Browse files like a folder, supports drag-and-drop of files or folders for one-click import.
- Smart Organization - AI can help organize folders and automatically categorize them.
📄 Preview Panel
- 9+ Format Preview - Supports formats like PDF, Word, Excel, PPT, code, Markdown, images, etc.
- Real-time Tracking + Editable - Automatically tracks file changes, supports real-time editing and debugging of Markdown, code, HTML.
🎨 AI Image Generation and Editing
- Smart Image Generation - Supports various image generation models such as Nano Banana (e.g.
gemini-3.1-flash-image,gemini-3-pro-image). - Image Recognition and Editing - AI-powered image analysis and editing features.
🌐 Multi-Channel Access
- WebUI Remote Access - Accessible via browser from any device on the network, supports mobile devices.
- Telegram Integration - Supports interaction via Telegram bots.
- Feishu Integration - Supports access and interaction via Feishu.
- Local Data Security - All data is stored in a local SQLite database, suitable for server deployment.
88API Integration Method
Parameter Entry
| AionUi field | Value |
|---|---|
| Model Platform | Custom |
| API Endpoint | https://api.88api.ai/v1 (/v1 is required) |
| Full URL | Off |
| API Key | The API Key created in the 88API console |
| Model Name | Pick from the dropdown; you can select several |
| Vision input, OpenAI request API | Keep the automatic option |
Create your API Key in the 88API console and keep it safe. Do not paste it into apps or pages from unknown sources.
One-click import: on the 88API console API Keys page, click Chat → AionUI next to the key you want to use. The browser opens AionUi with the New API platform, the address
https://api.88api.ai/v1and the API key pre-filled. Pick the models, keep the request protocol on OpenAI and confirm (when you pick aclaude-*model, AionUi switches the protocol to Anthropic automatically; change it back to OpenAI, otherwise requests go tohttps://api.88api.ai/v1/v1/messagesand fail with 404). You can also add it manually with the steps below.
Configuration Steps
-
Copy the API Key in 88API

-
Open AionUi Settings
- Go to the settings page in AionUi and open the Models page
- Click Add Model in the top-right corner
-
Fill in the platform information

- Model Platform: choose Custom
- API Endpoint: enter
https://api.88api.ai/v1(/v1is required) - Keep the Full URL switch off; AionUi appends the request path automatically
- API Key: paste the API Key copied from the 88API console
- Model Name: pick the models you need from the dropdown; you can select several (for example
gpt-6.1-solanddeepseek-v4.1-flash) - Keep Vision input on Automatic detection and OpenAI request API on Automatic (recommended)
- Click Confirm
-
Check that the models were added
The new models are listed under the Custom platform on the Models page. Models whose switch is on are ready to use.

-
Start Using
- Return to the chat page and select the Aion CLI assistant
- Select a configured 88API model in the top-right corner to start a conversation (AionUi currently uses the models added here only for Aion CLI; agents such as Claude Code and Codex use their own built-in configuration)
- Image input: select
deepseek-v4.1-flash, drag or paste an image into the input box and ask your question. Aion CLI reads the image with its ViewImage tool; if the reply says it cannot see the image, send the message again