swuecho/chat

What it solves

This project provides a self-hosted chat interface for interacting with various AI models, allowing users to manage conversations, organize history, and share their interactions. It solves the problem of needing a centralized, customizable chat UI that supports multiple model providers.

How it works

The system acts as a frontend and API layer that connects to LLM providers. It supports OpenAI, Claude, and locally hosted models via Ollama. Users can define system prompts, upload text or multimedia files, and and use a prompt management system with a '/' shortcut for quick access to saved prompts.

Who it’s for

Individuals or teams who want a private, self-hosted chat application with administrative controls, rate limiting, and the ability to organize and search through their AI conversation history.

Highlights

• Multi-model support: Compatible with OpenAI, Claude, and Ollama.
• Conversation Management: Creates shareable static pages (similar to ShareGPT) and provides a snapshot directory with full-text English search.
• Conversation Context: Automatically includes the latest 4 messages in the context window by default.
• Cusomizable Admin Controls: Includes an admin panel for rate limiting and selecting specific models for automatic title generation.
• File Support: Supports uploading text and multimedia files (depending on model capabilities).
• Mobile App: Includes a Flutter-based mobile application for chatting on the go.