ProductBiographic Intelligence
biographic-intelligence.metrodata.co
Biographic Intelligence is a full-stack analyst workstation for building structured psychological profiles from opt-in subjects’ narrative source material. Analysts ingest free-text reports through a React web app; a Cloudflare Workers pipeline stores raw documents in R2, queues LLM extraction into reviewable “atoms,” and runs contradiction, assessment, and hypothesis passes before facts are committed to versioned dossiers in D1. A human-in-the-loop review queue keeps analysts in control of what enters the record, while a dashboard surfaces structured intelligence and a subject-aware chat layer answers questions with citations back to reports and dossier content. The system is built as a TypeScript monorepo (shared Zod schemas, Hono API, Vite UI) on Cloudflare’s edge stack—Workers, D1, R2, and Queues—with Firebase Authentication for production sign-in.
Open product →
ProductNewspaper Publishing Management System
newspaper.metrodata.co
Newspaper Publishing Management System is a platform for personalized news: each user can run their own digital newspaper tuned to what they care about, from broad daily briefings to focused intelligence-style reports. You define the topics, tone, and depth you want; automated editorial pipelines research, draft, and refine stories on a schedule or on demand, then route them through review before they go live. A management console lets you shape multiple “papers,” adjust how content is produced, and control what gets published; separate reading experiences serve public editions and private, subscriber-only feeds. Built as a modern full-stack web application with a API backend, database, and browser-based admin and reader apps, it turns generalized news and niche analysis into a repeatable, personal publishing workflow rather than a one-off chat output.
Open product →
ProductMachine Agent Auction Protocol (MAAP)
MAAP API is a production-oriented backend for the Machine Agent Auction Protocol (MAAP)—a hardened auction system built for autonomous agents that buy and sell digital deliverables with on-chain USDC escrow. The service runs on Cloudflare Workers with D1, R2, and KV, and exposes a versioned REST API (documented in OpenAPI) for the full auction lifecycle: sellers pre-commit deliverables to object storage, bidders place funded bids with proxy/max-bid rules, scheduled jobs close auctions and settle winners, and an objection window plus admin dispute tools protect buyers and sellers. An integrated admin portal (Firebase-authenticated, role-based) supports operations, accounting, and configuration—including optional seller bonds—and companion tooling includes a load simulator and local cron ticker for realistic dev and testing. The stack uses TypeScript, Hono, and Vitest, with Solana funding flows stubbed locally and designed to plug into a real treasury and chain watcher in production.
ProductPatrol
patrol.metrodata.co
Patrol is a full-stack operations platform for private security firms, built as a TypeScript monorepo with an Express and PostgreSQL API, a React web portal for managers and clients, and a React Native (Expo) mobile app for field guards. Guards start shifts with background GPS tracking, scan QR-coded checkpoints within geofenced patrol windows, and file geotagged incident reports with photo and video attachments—even when offline, with automatic sync when connectivity returns. Managers monitor live guard positions and patrol compliance on an interactive map, review scan history, manage sites and checkpoints, and coordinate shift handoffs and branch notices; clients get a read-only view of activity and incidents at their assigned sites. The system supports multi-tenant security companies (branches, clients, and sites), role-based access via Firebase Authentication, and production-minded API safeguards including rate limiting and validated location and scan payloads.
Open product →
ProductLanguage Study System
Language Study System is a full-stack, AI-powered language learning platform built as a TypeScript monorepo: a React (Vite) dashboard talks to a Node.js (Fastify) API backed by PostgreSQL and OpenAI. Learners practice in a dual-column chat that shows their native and target languages side by side, with automatic translation, language detection, session transcripts, and dictionary lookup. At the core is a multi-agent Teaching Engine that runs structured lessons from a course plan—classifying each turn, tracking learner progress and placement (CEFR-style levels), and adapting objectives and difficulty over time—grounded in comprehensible-input principles rather than rote drills. The system supports anonymous study with optional Firebase account linking, plus rate limits and usage controls for safe public deployment. Stack highlights: React 19, Prisma, Docker, and an admin surface for tuning the teaching engine.
ProductGoal Compass
Goal Compass is a mobile personal productivity app that helps you work toward big goals more efficiently by turning vague ambitions into a clear map of what actually moves you forward. You build Goal Profiles that link your main objective to the everyday activities that support it, using ready-made templates or your own structure. Regular check-in surveys capture how well you’re doing on each activity; the app then scores your overall progress and highlights where you have the most untapped potential—the actions that could improve your goal score the most if you focused there next. Coaching insights, trend tracking, and visual views of your goal structure make it easier to see what’s working, what’s lagging, and where to put your energy; a diary lets you log related notes and experiences along the way. Whether you’re pursuing fitness, career, creative, or life-change goals, Goal Compass acts as a practical guide that points you toward the highest-impact next step instead of leaving you guessing where your effort should go.