25 tests, CI on Node 20 and 22

Open-source library for the part of an agent that must not be improvised: an LLM picks one of a known set of domains, and everything downstream comes from config it never sees.

A prompt injection in the input cannot name a handler that does not exist, and cannot waive a human approval. Zero dependencies.

316 applications submitted autonomously, 248 verified

Autonomous job application agent. Discovers listings, fills forms, runs QA — fully headless with Playwright + Chrome CDP.

17 job board APIs · CV tailored per posting against a fact registry · QA auditor pauses a worker after 3 failures and resumes it on a clean cycle.

21k lines of TypeScript, model routing by cost

Multi-agent orchestration dashboard. Work is classified by an LLM, routed by deterministic config, and held at a human approval gate before anything executes.

Claude Agent SDK · Chrome CDP · Telegram API · Groq intent classification · 6 scheduler jobs

27k lines of TypeScript, live in production

Browser strategy game with AI-driven NPC dialogs and real-time physics. Vertical slice: Operator League.

React Three Fiber + Rapier physics · Groq SDK for in-game dialogs · Supabase realtime matchmaking

37k lines, built solo end to end

B2B SaaS that generates e-commerce product listings for Amazon and Etsy from an LLM pipeline, with per-marketplace constraints validated before anything is shown.

Auth, subscription tiers, Stripe billing and row level security. Built solo, front to back.

RAG over a standards corpus, key rotation on rate limits

Enterprise document reconciliation platform. Semantic engine detects inconsistencies and duplicates across large document sets.

Private deployment for enterprise client · Railway backend + Vercel frontend · semantic similarity engine · Python + Next.js

LLMs & AI

Frontend

Backend & DB

AI Infra & Tools