WebFlow OS

An AI-powered workspace combining autonomous browsing, deep research, real-time voice interactions, and emotion-aware responses.

Comment

WebFlow OS

An AI-powered workspace combining autonomous browsing, deep research, real-time voice interactions, and emotion-aware responses.

Inspiration

Modern workflows are fragmented across browsers, research tools, voice assistants, and automation platforms. We wanted to build a single intelligent operating layer for the web — one that can browse, research, listen, reason, and adapt emotionally in real time.

WebFlow OS explores what happens when AI becomes an active collaborator, not just a chatbot.

What it does

WebFlow OS acts as an autonomous AI workspace capable of:

  • Browsing websites and completing tasks automatically
  • Performing deep multi-source research
  • Holding real-time voice conversations
  • Detecting emotional tone and adapting responses
  • Extracting structured information from the web
  • Maintaining persistent conversational memory
  • Understanding screen context through sharing

Instead of switching tools, users interact with one intelligent system that executes workflows end-to-end.

Features

Core Capabilities

  • Yutori Browsing — Cloud browser automation that navigates websites, extracts data, and completes tasks autonomously
  • Research Tasks — Deep multi-source research with comprehensive reports and analysis
  • Tavily Search — AI-optimized web search with intelligent filtering
  • Tavily Extract — Structured webpage parsing for AI reasoning
  • Emotion Detection — Real-time emotion analysis powered by Modulate Velma-2
  • Real-Time Voice — Streaming voice conversations with live transcription
  • Neo4j Persistent Memory — Context-aware AI interactions across sessions using the Neo4j memory agent
  • Screen Sharing — Visual collaboration and contextual understanding

Use Cases

  • Market Research — Automated competitor analysis and trend monitoring
  • Lead Generation — Extract contacts and company intelligence
  • Content Curation — Discover and summarize web content
  • Price Monitoring — Track e-commerce pricing automatically
  • Academic Research — Multi-source analytical reports
  • Customer Support — Emotion-aware adaptive responses

Who Is This For?

  • Startup Founders — Automated research without extra hires
  • Sales Teams — Prospect intelligence and preparation
  • Content Creators — Faster topic discovery and sourcing
  • Researchers — Deep analysis across the web
  • Developers — AI-powered browser workflows
  • Product Managers — Competitive and user insight analysis

Tech Stack

  • Built with tech

  • Frontend: React + Vite

  • Backend: Python (FastAPI)

  • Voice AI: Google Gemini Live API / ADK

  • Emotion Detection: Modulate Velma-2 STT

  • Web Search: Tavily API

  • Browser Automation: Yutori API, Playwright

Project Structure

├── agents/
│   └── app/
│       ├── frontend/
│       ├── chat_agent/
│       ├── chatkit_integration/
│       ├── tavily_tools_adk.py
│       ├── yutori_tools.py
│       ├── velma_stt_service.py
│       ├── browser_service.py
│       └── main.py
└── README.md

Built With

Share this project:

Updates