WebFlow OS – Agentic Web Workspace
A solo build for the Autonomous Agents Hackathon. Modern workflows are fragmented across browsers, research tools, voice assistants, and automation platforms. WebFlow OS is one workspace that browses websites autonomously, runs multi-source research, holds real-time voice conversations, detects emotional tone in speech, extracts structured data from pages, and remembers context across sessions in a graph.
Project Features
Autonomous Browsing
Yutori and Playwright agents open sites, follow flows, and pull structured data on request, including price monitoring across e-commerce sites.
Multi-Source Research
Tavily search results are gathered, cross-checked, and summarized into one answer instead of ten tabs.
Voice With Emotional Awareness
Gemini Live handles the conversation while Velma-2 tags the speaker's tone, so the assistant can adjust when a user is frustrated.
Graph Memory
Neo4j keeps conversational context and extracted facts across sessions, so yesterday's research is today's starting point.
Project Info
02/2026
Sole Developer
Solo Project
Hackathon Submission
Technologies Used
- • React, Vite
- • Python, FastAPI
- • Gemini Live API
- • Modulate Velma-2 STT
- • Tavily API
- • Yutori API, Playwright
- • Neo4j
Project Gallery
Interested in This Project?
I'm always excited to discuss my projects and share insights about the development process. Feel free to reach out if you'd like to know more!