This project showcases how to build practical AI agents with CopilotKit, Google DeepMind’s Gemini, and LangGraph.
It includes two agents, exposed through a Next.js frontend and a FastAPI backend.

✨ Features

Post Generator Agent
Generate LinkedIn and Twitter posts from the context you provide.
Useful for creating professional, context-aware social content.
Stack Analyzer Agent
Provide a URL and get a detailed breakdown of the site’s technology stack.
Quickly identify frameworks, libraries, and infrastructure used.

🛠️ Tech Stack

Frontend: Next.js
Backend: FastAPI
Agents: Google Gemini + LangGraph
UI Layer: CopilotKit

📌 About

This demo illustrates how CopilotKit can be paired with LangGraph and Gemini to create agents that are:

Context-aware (understand the input you provide)
Task-focused (generate content or analyze stacks)
UI-integrated (feels like part of your app, not just a chatbox)

Project Structure

/ — Next.js 15 app (UI) in the Project Root
agent/ — FastAPI backend agent (Python)

🚀 Getting Started

1. Clone the repository

Clone this repo git clone <project URL>

2. Environment Configuration

Create a .env file in each relevant directory as needed.

Backend (`agent/.env`):

GOOGLE_API_KEY=<<your-gemini-key-here>>

Frontend (`/.env`):

GOOGLE_API_KEY=<<your-gemini-key-here>>

3. Running the project

pnpm install
pnpm dev

Open http://localhost:3000 in your browser to view the app.

Notes

Ensure the backend agent is running before using the frontend.
Update environment variables as needed for your deployment.

README.md Unescape Escape

Open Gemini Canvas

CopilotKit + Google DeepMind (Gemini) + LangGraph Template