Adding project minimal-ts-agent

[Pash10g]

This pull request introduces a new minimal Retrieval-Augmented Generation (RAG) agent demo app in TypeScript, demonstrating how to build an AI agent that answers questions about the MongoDB Brand Book using MongoDB Atlas Vector Search, Voyage AI embeddings, and the Vercel AI SDK's ToolLoopAgent. The changes set up the project structure, dependencies, configuration, data ingestion, API routes, UI, and styling.

The most important changes are:

Project Setup and Configuration

• Added initial project files for the minimal RAG agent demo, including package.json with dependencies for Next.js, Vercel AI SDK, MongoDB, Voyage AI, Tailwind CSS, and TypeScript. Also included example environment variables in .env.local.example and a comprehensive .gitignore for Node/Next.js projects. [1] [2] [3]
• Added ESLint and PostCSS configuration files for code quality and styling support. [1] [2]
• Added a basic Next.js config file (next.config.ts).

Documentation

• Added a detailed README.md explaining the project purpose, architecture, tech stack, setup instructions, and agent workflow.

Data Ingestion and Vector Search

• Added scripts/ingest.ts to embed MongoDB Brand Book sections using Voyage AI, insert them into MongoDB Atlas, and create a vector search index for RAG.

API and Agent Integration

• Implemented the /api/chat endpoint (src/app/api/chat/route.ts) to handle chat requests, invoke the agent, and stream responses.

UI and Styling

• Set up global styles (globals.css) with Tailwind CSS and custom theme variables for a branded dark UI.
• Added a root layout (layout.tsx) that loads custom fonts and applies consistent styling and metadata.

Wrike: https://www.wrike.com/open.htm?id=4370041280