VisionForge AI

Multimodal AI agent by Sharad Khare.

VisionForge AI analyzes images and videos with Google Gemini for reasoning, Q&A, and practical insights.

Features

Image reasoning with Gemini 2.5 Pro (app_reasoning.py)
Video analysis with Gemini 2.5 Flash (app.py)
Support for JPG, PNG, MP4, MOV, and AVI
Streamlit interface with secure sidebar API key input
Modular services for image and video workflows

Quick start

cd visionforge-ai
pip install -r requirements.txt
streamlit run app.py

Image reasoning mode:

streamlit run app_reasoning.py

Required API key

Google Gemini API key from Google AI Studio

How it works

1. Upload an image or video file.

2. Ask a task-specific question or analysis prompt.

3. VisionForge sends the media to Gemini through an Agno agent.

4. You receive a markdown-formatted response in the app.

Project structure

visionforge-ai/
├── app.py
├── app_reasoning.py
├── visionforge/
│   ├── config.py
│   └── services.py
├── requirements.txt
├── pyproject.toml
└── README.md

Topics

Sharad Kharemultimodal AI agentGemini video analysisimage reasoning AIvision language modelStreamlit multimodal appvideo Q&A agentGoogle Gemini agentvisual understanding AI.

Who this is for

Teams building multimodal AI assistants
Developers experimenting with Gemini vision workflows
Educators creating visual reasoning demos
Product builders shipping image/video intelligence features

Use cases

Analyze uploaded videos and answer context-specific questions
Perform image reasoning for object, scene, and task insights
Prototype visual Q&A assistants with Streamlit
Build Gemini-powered multimodal application demos

Author

Website: sharadkhare.in

License

Loading file tree…

Select a file to preview its source.

bash

cd visionforge-ai
pip install -r requirements.txt
streamlit run app.py