Report Abuse

Basic Information

Better Chatbot is an open-source chatbot template and developer-focused platform for building modern AI chat interfaces. It is built with Next.js and the Vercel AI SDK and is designed to integrate external tools using the Model Context Protocol (MCP). The repository provides a working chat application plus infrastructure to add and call MCP tools, publish visual workflows as callable tools, and wire LLM responses to automated multi-step tool executions. It supports realtime voice interactions via an LLM Realtime API, local or Docker-based hosting, and production deployment on Vercel. The project expects developer setup with LLM provider API keys and optional PostgreSQL for persistence. Included guides cover MCP server setup, Docker hosting, OAuth, system prompts, and adding OpenAI-like providers. The codebase serves both as a deployable chatbot and as a framework for composing tool-backed agentic workflows.

Links

App Details

Features
MCP integration to discover and call external tools from chat, plus support for multiple MCP servers and presets. Visual workflow authoring that turns LLM/Tool sequences into reusable @workflow tools. Browser automation example using a playwright-mcp tool to let the model control web actions. Realtime voice assistant demo built on a Realtime LLM API combined with MCP tool execution. Quick tool mentions using @ to bind tools for single responses, tool selection/presets for persistent tool availability, and a Tool Choice Mode (Auto/Manual/None) to control autonomy. Built-in tools include semantic web search, a JS executor, HTTP client, interactive data tables, and chart generation. Deployment guides for Docker Compose and Vercel, PostgreSQL integration, and OAuth sign-in configuration are provided.
Use Cases
This repository helps developers and teams prototype and deploy tool-enabled chat assistants quickly. It provides a ready-made UI and backend patterns to connect multiple LLM providers, manage API keys, and persist data in PostgreSQL while exposing MCP tools so chat models can perform real-world actions. Visual workflows let non-programmers and engineers convert multi-step sequences into callable tools, speeding automation and reuse. Tool mention and selection features reduce token usage and let you control which tools are visible to the model. The JS executor and data visualization tools support interactive computation and reporting inside chats. Comprehensive guides, Docker and Vercel deployment recipes, and community support make it practical for local development, staging, and production use.

Please fill the required fields*