ErgoSafe Reborn

ErgoSafe Reborn: A Gemini Stewardship Engine for OHS compliance. Melly prevents Repetitive Stress Injury and Sec 37/38 liability via AI Cognitive Handshakes and automated legal safety dossiers.

Comment

InspirationIn : South Africa, Occupational Health and Safety (OHS) is often a "paper exercise" until tragedy strikes. Under the OHS Act 85 of 1993, executives face strict personal liability for workplace injuries. With remote work and distributed industrial teams, tracking ergonomic rot and cognitive fatigue is nearly impossible. We were inspired to transform the dense legal requirements of Sections 37 and 38 into a living, breathing Digital Wingman that protects both the worker's health and the CEO's legal standing.

What it does:

ErgoSafe Reborn is a Stewardship & Fatigue Engine that shields executives from the strict vicarious liability of Sections 37 and 38 of the South African OHS Act. It acts as a "Digital Wingman" for remote and industrial workforces, performing mandatory Cognitive Handshakes before login to establish safety baselines. Our AI agent, Melly, detects ergonomic deterioration and cognitive fatigue in real-time, issuing "Professional Resets" and localized Zulu/English First Aid interventions to prevent accidents before they occur.

How we built it: We architected ErgoSafe Reborn as a Rapid Agent ecosystem:

Orchestration: Google Cloud Agent Builder manages the agentic reasoning and intent classification for Melly.The Intelligence: Gemini powers the context-aware safety coaching and multilingual support. Partner Integration: We utilized the MongoDB MCP Server to create a live telemetry ledger. Every safety scan and compliance handshake is embedded into MongoDB, allowing the agent to query real-time data for "DOA Lockout" decisions. CI/CD: The frontend is built with React and Tailwind CSS, deployed via Vercel with a seamless GitHub integration.

Challenges we ran into: Engineering an enterprise-grade safety engine on an i5 ASUS with 8GB RAM forced us to prioritize extreme code efficiency. We faced significant challenges in:

Low-Latency Agentics: Optimizing the Google Cloud ADK to ensure Melly could provide life-saving First Aid instructions without lag on constrained hardware.The "Spy-Tool" Paradox: Balancing deep postural tracking with a "Privacy-First" philosophy, ensuring data is used for coaching rather than punitive surveillance.Legal Logic Mapping: Translating complex statutory requirements from the OHS Act of 1993 into actionable "If-This-Then-Agent" logic.

Accomplishments that we're proud of: We successfully built a Zero-Knowledge Dossier system that automates legal compliance for corporate boards. We are incredibly proud of our multilingual inclusivity—Melly’s ability to provide critical first-aid protocols in Zulu ensures that safety is accessible to the entire South African workforce. Earning our GDG Discovery Badge during the process served as proof that we could bridge the gap between "Idea" and "Rapid Agent."

What we learned: We mastered the art of Agentic Stewardship. We learned that AI shouldn't just answer questions; it should act as an active guardian. Through the Google Cloud Innovator labs, we discovered how to use MongoDB MCP to give an AI agent a "long-term memory," allowing Melly to track fatigue trends over days, not just seconds.

What's next for ErgoSafe Reborn: The next phase is Predictive Ergonomics. We are moving toward integrating computer vision to detect "Bed-Working" and "Slouch-Fatigue" before the user even feels the physical strain. We plan to expand our legal mapping to cover SADC regional safety standards, turning ErgoSafe Reborn into the definitive safety shield for the entire African industrial sector.

Built With

  • antigravity
  • gemini-api
  • google-cloud-agent-builder
  • mongodb-mcp
  • next.js
  • node.js
  • react
  • tailwind-css
  • typescript
  • vercel
  • vertex-ai
Share this project:

Updates