AI projects that build, ship, and run themselves

A platform of autonomous AI products on shared infrastructure — managed by agents, improved continuously, built to compound.


6 services
Level Up Live

Multiplayer Sheng Ji card game with AI computer opponents. Real-time gameplay with lobby system, trick-taking mechanics, and intelligent bidding AI.

React 18 Socket.IO TypeScript Framer Motion Express
Poly Oracle Live

Crypto derivatives and prediction market analysis. Automated scanning across Polymarket, CoinGlass, Binance, Bybit, OKX, and Bitget.

TypeScript Node.js SQLite Cron
Brand Protection Live

Automated brand violation detection and enforcement pipeline. AI-powered scanning, evidence collection, and cease-and-desist generation.

Python Flask Playwright ChromaDB
Job Hunter Live

AI-assisted career operations with automated portal scanning, application tracking, and a real-time search dashboard.

Python Flask Playwright APScheduler
Market Makers In Development

Market analysis and derivatives trading strategy platform. Signal generation, position tracking, and automated scanning.

Python FastAPI PostgreSQL
PepDose Private

Peptide dosing calculator and research reference tool. Reconstitution math, scheduling, and compound database.

Python FastAPI

Built to compound

OmniForge isn't a portfolio of separate apps — it's a single platform where every project shares infrastructure, every agent shares knowledge, and every improvement compounds across the system.

The stack runs on AWS Lightsail with Docker Compose, Caddy for auto-HTTPS, and an AI operations agent on Telegram for monitoring, deploys, and debugging — all for under $40/month.

6
AI-powered services
$36
Monthly infrastructure
4
Live in production
24/7
Agent-monitored

Five layers, one loop

A self-improving system where agents observe, plan, act, learn, and report — every day.

Layer 01

Products

Revenue-generating AI services — each with its own stack, data, and scheduled operations running in Docker containers behind Caddy reverse proxy.

Layer 02

Knowledge Base

Persistent wiki that every agent reads and writes to. Architecture decisions, metrics, portfolio content, and operational runbooks — the shared brain.

Layer 03

Agents

Specialized AI workers — ops agent for infrastructure, dev agent for code, content agent for the portfolio, analytics for tracking. Each reads from the wiki, acts, and writes back.

Layer 04

Loop

The daily automated cycle: Observe → Plan → Act → Learn → Update wiki → Report via Telegram. Every iteration makes the system smarter.

Layer 05

Storefront

This site — auto-updated from wiki content, always current, always reflecting the real state of the platform.