What is AIME-BOX?

AIME-BOX is a multi-platform desktop chat client built with LangChain and Electron. It supports local knowledge bases, tool calling, and multiple intelligent agents, aiming for fully offline local execution.

How to use AIME-BOX?

Add a model provider (e.g., Ollama), then manage and enable the desired models. Start chatting directly. Download local models for features like knowledge base retrieval, OCR, speech recognition, and speech synthesis. Use npm run start for development and npm run package for Windows builds.

Key features of AIME-BOX

- Multiple model provider support: Ollama, OpenAI, DeepSeek, Anthropic, etc.
- Tool calling: web loader, web search, STT, TTS, OCR.
- Local knowledge base engine using LanceDB with image, docx, doc, text, URL inputs.
- Multi‑agent system: data extraction, code assistant, translation assistant.
- Prompt saving functionality.
- MCP (Model Context Protocol) support for model integration.

Use cases of AIME-BOX

- Extract structured data from files or folders based on user‑specified fields.
- Generate and execute Python scripts in a virtual environment for automation tasks.
- Chat with local LLMs using a variety of tools for search, image analysis, and speech.
- Build a fully offline assistant with local knowledge retrieval and voice interaction.

FAQ from AIME-BOX

Which AI models does AIME-BOX support?

It supports Ollama, OpenAI, DeepSeek, Anthropic, and other providers. Local models can be managed for offline use.

Does AIME-BOX support MCP servers?

Yes, MCP is supported and can be connected for model tool calls.

What is the license of AIME-BOX?

It is licensed under the Apache License 2.0.

Can all features run fully offline?

The project aims for fully offline local execution, but some features (e.g., model downloads) may require an initial internet connection.

What are current limitations?

Multilingual support and document chat (ChatDoc‑like) are not yet implemented. Speech and vision features depend on locally downloaded models.