FranciscoMoretti/chat-js

What it solves

ChatJS provides a pre-built foundation for AI chat applications, eliminating the need for developers to repeatedly build common infrastructure like authentication, model integration, and streaming from scratch.

How it works

It is a full-stack framework built with Next.js and TypeScript that integrates several key technologies:

• AI Gateway: Provides unified access to over 120 models from providers like Claude, GPT, and Gemini.
• EVE Runtime: A native runtime that manages durable conversation transcripts, execution state, approvals, and stream recovery.
• AI SDK: Utilizes the Vercel AI SDK for TypeScript-based AI toolkit functionality.
• Infrastructure: Uses PostgreSQL via Drizzle ORM for data, Vercel Blob for storage, and Better Auth for user management.

Who it’s for

Developers who want to quickly launch a professional AI chat application with advanced features (like branching conversations and code execution) without spending time on the boilerplate infrastructure.

Highlights

• Extensive Model Support: Access to 120+ models via a single API.
• Agentic Capabilities: Includes support for the Model Context Protocol (MCP), real-time web search, and sandboxed code execution.
• Conversation Management: Supports branching (forking) conversations and sharing them via public links.
• Rich Media: Built-in support for drag-and-drop attachments including PDFs and images, as well as AI image generation.
• Multi-platform: Can be packaged as a native desktop app for macOS, Windows, and Linux using Electron.