🚀 Fission AI

Fission AI is a powerful, multimodal AI platform that combines state-of-the-art models and tools to transform text into images, videos, research reports, and more. It is designed for creators, educators, developers, and researchers who want seamless AI-powered content generation — all from a single interface.

---

🔮 Key Features

• 🖼️ Text to Image
  Generate up to 5 high-quality images using Stable Diffusion.
  Captions generated with BLIP.

• 🎬 Text to Video
  Automatically stitches images into a video with generated narration using gTTS.

• 💬 Text to Text
  Uses Gemma 3B (via Ollama) to generate or rewrite text with LLM capabilities.

• 📚 Text to Deep Research

  • Extracts keywords
  • Searches the web using DuckDuckGo API
  • Summarizes findings using Gemma 3B
  • Generates a downloadable PDF report

---

🧠 Tech Stack

Category	Tools / Frameworks
Backend	Flask, Python, ngrok, Google Colab
Frontend	Streamlit
Machine Learning	PyTorch, BLIP, Stable Diffusion, Gemma 3B
Text-to-Speech	gTTS
LLM Integration	Ollama (Gemma 3B)
Web Search	DuckDuckGo API
Video & Audio Tools	FFmpeg, MoviePy
IDE / Development	VS Code
Hosting / Tunnel	ngrok

---

🏗️ System Architecture

User (Streamlit App)
      │
      ▼
[Frontend: Streamlit] ─────────┐
                               ▼
        [Backend: Flask server (ngrok tunnel)]
                               │
           ┌────────────┬──────┼───────────┬────────────┐
           ▼            ▼      ▼           ▼            ▼
[Stable Diffusion] [BLIP] [gTTS] [Gemma 3B] [DuckDuckGo API]
                               │
                          [Google Colab Runtime]
                               │
                         [Returns ZIP / PDF / MP4]

---

⚙️ Installation & Setup

⚠️ Note: This project uses Google Colab as the backend runtime, tunneled using ngrok.

1. Clone the repository

git clone https://github.com/avarshvir/fission_ai.git
cd fission_ai

2. Install dependencies

pip install -r requirements.txt

3. Set up ngrok

- get ngrok token after setting your account on ngrok.
- add your token 
    - NGROK_AUTH_TOKEN = "your ngrok token" in main_model.ipynb google colab
- run the all cell of main_model.ipynb
- get ngrok url from colab output and place the url in main.py file that act as backend
    - NGROK_URL = "ngrok_url" 

4. Make Google Drive Link Public

- create a folder name "GeneratedImages" inside mydrive of google drive.
- make "GeneratedImages" folder public and copy the link.
- place public link in main.py file gdrive_link = "public drive link" 2 times.

5. Launch the Streamlit frontend

after cd fission_ai
cd frontend
streamlit run main.py

6. Choose your option and see results

Text-to-Image
Text-to-Video
Text-to-Text
Text-to Deep Research

---

⚠️ Note:

make sure you download dependencies, setup your ngrok account, 
public your "GeneratedImages" folder of Google Drive, 
and choose T4 GPU of Google Colab.

---

📦 Folder Structure

fission_ai/
│
├── assets
|   ├── output
|   ├── final_narrated_video   # Stitch images/audio to final video
|   └── output.zip             # Stitch images/audio to final video
|
├── backend/
│   └── backend.txt
|
├── data/
│   └── dataset_links.txt
│
├── frontend/
│   ├── images
|   ├── resized_images
|   ├── videos
|   ├── keyword_search.py
|   ├── main.py               # Actual Main file
|   ├── summarizer.py
|   └── text_query.py
|
├── model/
│   ├── main_model.ipynb      # main file for google colab
|   ├── model.txt
|   └── ok8.ipynb
|
├── CODE_OF_CONDUCT.md
├── CONTRIBUTING.md
├── LICENSE
├── NOTICE
├── README.md
└── requirements.txt

---

💡 Future Plans

• Add user authentication & plans (Free, Pro, AI+, Enterprise)

• Provide cloud-based persistent storage

• Rate-limited access to premium APIs

• Launch SaaS version under Jaiho Labs (Subsidiary of Jaiho Digital)

---

🤝 Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repo

2. Create your feature branch (git checkout -b feature-name)

3. Commit your changes (git commit -am 'Add new feature')

4. Push to the branch (git push origin feature-name)

5. Open a pull request

---

📬 Contact

• Author: Arshvir

---

🌟 Star this repo to support the project! 😊