A newer version of the Gradio SDK is available: 6.30.0
title: VLA Demo - Codevally AI
emoji: 🤖
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 5.25.0
app_file: app.py
pinned: false
license: other
short_description: Vision Language Action pipeline — open-vocabulary AI demo
🤖 Vision Language Action (VLA) Demo
Built by Codevally — AI-Powered Software Development & AI Solutions
Upload an image and enter a natural-language command. The system detects objects using Grounding DINO (open-vocabulary), interprets your instruction with an LLM, generates a structured action plan, and visualizes the result.
Try it: "Pick up the defective board and place it in the red bin"
How it works
- 📸 Vision — Grounding DINO detects objects based on your command text
- 🧠 Language — LLM (OpenAI / Groq) interprets the command
- 📋 Planning — Structured action steps are generated
- 🎬 Action — Annotated visualization is produced
🏢 About Codevally
Codevally is a forward-thinking technology company at the intersection of AI and software development. We specialize in building intelligent systems, AI-powered applications, and custom AI solutions that transform businesses — from traditional software to cutting-edge machine learning implementations.
🔬 Our AI Specializations
| Domain | Capabilities |
|---|---|
| 🖼️ Computer Vision | Object detection, image classification, video analytics, visual search |
| 🤖 Language Models (LLMs) | OpenAI, Claude, Gemini, Llama fine-tuning, RAG systems, AI agents |
| 💬 NLP | Sentiment analysis, NER, text summarization, question answering |
| 📈 Predictive AI | Forecasting, anomaly detection, churn & risk scoring |
✅ Why Codevally
- AI-First Approach — Deep expertise in AI/ML and modern LLM technologies
- End-to-End Solutions — From concept to deployment and monitoring (MLOps)
- Practical AI — Focused on real business value, not just technology
- Responsible AI — Ethics, bias detection, and governance built-in
This demo is an open-vocabulary Vision Language Action pipeline built by the Codevally AI team. Learn more →