AlwaysLateToEverything

The “VoxVisual” tool allows users to perform hands-free data analysis. It transforms spoken queries into high-fidelity, animated SVG visualizations. Unlike static dashboards, the UI is ephemeral and generated on-demand, powered entirely by the Anthropic Claude API for intelligence and the Redis Agent Memory Server to maintain a dual-tier “conversational memory” of the data exploration — ephemeral working memory for the current session and persistent long-term memory for cross-session intelligence. The proof-of-concept ships with a built-in demo dataset: PedalForce Bicycles 2025 sales data, stored as JSON in Redis.

for Technologies, Frameworks, and Libraries:
Anthropic Claude API - Core reasoning engine that generates real-time SVG UI from voice prompts
Redis - Manages working + long-term memory for cross-session intelligence and real-time dataset storage + structured querying
Next.js - Full-stack React framework for API + frontend integration
React - Dynamic SVG injection + state-driven rendering
Tailwind CSS Responsive styling system
SVG + CSS Animations - AI-generated responsive visualizations
Web Speech API - Voice input capture
SpeechSynthesis API - AI voice output
Python - Tool execution + Claude orchestration
Docker - Deployment

I worked on previous claude code project and Dedalus, called DebtReversionAI, another project called “Google Meta Agent”, and another called “MediaSynth”

All my projects can be seen at my Github address:
https://github.com/ithllc