LoLLMs WebUI

LoLLMs WebUI (originally released as gpt4all-ui) is an open-source hub for developers and power users who want one interface to access many large language and multimodal AI models locally, instead of juggling separate tools for text, image, and music generation. It supports tasks such as writing, coding, data organization, and image or music generation.

A user installs the project, selects a binding and model, such as a Hugging Face local model, a GGUF/GGML local model, or a remote service like Ollama, OpenAI, or Anthropic, then picks a predefined 'personality' with its own welcome message for a task. The project is run locally via an installation script or manual setup, then started with app.py to open the web interface.

It offers over 500 AI expert personalities and more than 2,500 fine-tuned models across domains, image generation through services like Stable Diffusion, Flux, ComfyUI, DALL-E, or Midjourney, video generation through services like RunwayML or CogVideoX, and music generation through MusicGen. A Smart Routing feature can select among configured models to optimize generation for cost or speed, and discussions are stored in a local database that can be searched, exported, or deleted.

LoLLMs WebUI is open source under the Apache-2.0 license, supports Docker, conda, or manual virtual environment setup, and has more than 4.8k GitHub stars. The project's maintainer notes that a newer, more advanced multi-user project called lollms is in development, and that this webui will eventually be replaced by it, with only minimal support maintained in the meantime.

Use cases

• Getting coding assistance in Python, C++, or JavaScript
• Generating images, video, or music from a single interface