Alpaca WebUI

Alpaca WebUI, initially crafted for Ollama, is a chat conversation interface featuring markup formatting and code syntax highlighting. It supports a variety of LLM endpoints through the OpenAI Chat Completions API and now includes a RAG (Retrieval-Augmented Generation) feature, allowing users to engage in conversations with information pulled from uploaded documents.

The app lets you easily try out different AI models from various services to find answers using your documents. It simplifies mixing and matching these models to see which one works best for your needs.

Has

• Support for your local or remote Ollama server
• Compatible with OpenAI Chat Completions API
• Works with OpenAI, Together.ai, Mistral.ai, Groq.com and Ollama
• Upload document (pdf, txt, docx) for RAG usage
• Modify System Prompt for Chat and RAG
• Markup Formatting
• Code Syntax Highlighting
• List available models for services supporting OpenAI API model endpoint (and listing of Ollama models)
• Mobile-Friendly and Desktop-Friendly Layout
• Generate images with OpenAI dall-e-3 and dall-e-2 models (write /image followed by prompt in chat)

Pending Implementation

• Modify Parameters (like temperature, mirostat ...)
• Attach ImageFile
• Chat History

Env

This is a Next.js project bootstrapped with create-next-app.

Getting Started

Run the development server:

1. git clone https://github.com/mmo80/alpaca-webui.git
2. Setup and install your Weaviate database or just use the docker-compose weaviate file in the repository with the following command docker-compose -f docker-compose.weaviate.yaml up --build --detach.
3. Check the environment variables in .env.development
4. Then just run npm install followed by npm run dev 😎

Open http://localhost:3000 with your browser to start using the app.

Docker Compose

This command sets up two containers: one for the app and another for the Weaviate vector database. (Please note that there is no data persistence on a mapped volume in this configuration.)

docker compose up --detach

Then Open http://localhost:3033 with your browser to start using the app.

The official Alpaca WebUI Docker Image forloopse/alpaca-webui is available on Docker Hub.

YouTube Overview