Piko.ai

Your Friendly Neighbourhood Learning Assistant

Piko.ai is an AI-powered learning assistant that turns any topic into beautiful Manim-based animations and concise explanations — all from a simple prompt.

🚀 Features

• 🧠 AI-Powered Animation Generation

  Enter a topic (e.g. “binary trees” or “rotational motion”), and instantly generate Manim code using Google's Gemini LLM.

• 🎬 Automatic Video Compilation

  The generated Python code is saved, compiled using Manim CLI, and the output .mp4 video is served automatically to the frontend.

• 📝 Topic Explanation with LLM

  Along with the animation, a concise textual explanation is generated to help users understand the topic better.

• 🎧 Audio Narration

  Support for accompanying .wav audio narration for enhanced understanding.

• 🌐 Modern Frontend Experience

  Built with Next.js, Tailwind CSS, Framer Motion, and ShadCN for smooth UX, animations, and mobile responsiveness.

🖼️ Preview

Landing Page	Generated Animation + Explanation

For a demo video, click here.

🛠️ Tech Stack

🔹 Frontend

• Next.js 14 with App Router
• Tailwind CSS for styling
• Framer Motion for animations
• ShadCN UI for beautiful components

🔹 Backend

• FastAPI for API endpoints and routing
• Google Gemini API for code + explanation generation
• Manim CLI for compiling Python animations
• FAISS for creating a vector index of documentation

---

🧪 How It Works

Mermaid Block:

graph TD;
    User["User enters a topic"] -->|"POST /code/{topic}"| FastAPI
    FastAPI -->|"Generate search query from topic"| QueryGen
    QueryGen -->|"Perform FAISS similarity search"| VectorDB["Vector Store (FAISS)"]
    VectorDB -->|"Top docs as context"| FastAPI
    FastAPI -->|"Prompt Gemini with context + topic"| GeminiAPI
    GeminiAPI -->|"Returns Manim code + explanation"| FastAPI
    FastAPI -->|"Compile with Manim CLI"| MP4[".mp4 Video"]
    FastAPI -->|"Return video + explanation"| Frontend
    Frontend -->|"Display animation and text"| User

Loading

Running Locally

1. Clone & Setup Backend

git clone https://github.com/yourusername/piko-ai.git
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt

2. Run FastAPI server

uvicorn main:app --host 0.0.0.0 --port 8000

3. Setup Frontend

cd ../frontend
npm install
npm run dev

Environment Variables

Create a .env file in the root folder:

GEMINI_API_KEY = your_google_gemini_api_key