Community platforms, AI agents, GHL integrations, and internal tools — built end-to-end by one senior engineer commanding a fleet of AI agents. You get agency-grade software at a fraction of the timeline, and you own the code.
A complete community + course platform — auth with magic links, role-based access, lesson management, file uploads, progress tracking, learner submissions, comments, and admin analytics. Everything Skool does, except you own it: no $99/mo, no per-transaction fees, your brand, your data.
Case study +Interactive explorer for all 413+ GoHighLevel API v2 endpoints. Fire real calls, get AI-powered guidance, export to cURL, Make.com, and n8n. The fastest way to build on GHL.
Case study +AI conversation engine that qualifies leads and drives them to booking — covering ~80% of what Closebot does, fully custom and self-hosted. No per-seat SaaS pricing, tuned to your scripts and offers.
Case study +An education platform plus the engine behind my delivery speed: the App Builder takes an idea through survey → spec → scaffold → production phases. It's how one builder ships like a team.
Case study +AI voice agents that answer, qualify, and route calls — natural conversation, call recording, transcription, and CRM sync. Your front desk, around the clock.
Case study +An AI that works leads over SMS and chat, handles objections, and books straight into GHL calendars — checking real availability, confirming, and reducing no-shows with smart follow-up.
Case study +A client-facing program delivery platform built for a paid engagement — structured curriculum, member progress, dashboards, and accountability tooling that keeps clients engaged and renewing.
Case study +Snap-and-render invoicing: generate clean, branded invoices in seconds instead of fighting templates. A small tool that shows the philosophy — automate the boring, keep it beautiful.
Case study +Full Claude setup for teams: MCP servers, custom skills, agent workflows, and Cowork rollouts — turning an AI subscription into an operating system for the business. The same system that powers my own delivery speed, installed in yours.
Case study +Conversation IQ analytics · KPI tracker · Mind-mapping tools · GHL toolkit MCP server · AI front-desk HQ — 40+ repos and counting.
Every feature below already ships in a live production app — not a roadmap, not a demo. Order one as a standalone module, bundle several into a build, or fold them into a care retainer. Same stack every time: React + TypeScript on Supabase, with row-level security and edge functions.
The conversation layer: receptionists, qualifiers, and assistants that answer, book, and get smarter on their own.
Talk to the CRM you already run on — directly, not through duct-taped Zaps.
Where every build starts — the security posture under all of it.
Take money and provision access the moment it lands.
The retention layer GHL is weak on — curriculum your clients actually finish.
A real community you own — not a dead Facebook group or rented Skool.
Events, reminders, and booking — synced to the calendar you live in.
The machinery that keeps clients engaged — and flags the ones about to slip.
The engine behind the delivery speed — and the gates that keep quality high.
The control room — see what's happening and let prospects try before they buy.
// Fixed scope · one revision round included · you own the code
A structured survey turns your idea into a complete build spec — data model, screens, integrations, edge cases — before a line of code is written. No scope drift, no surprises.
AI agents handle scaffolding, components, and integration code on a battle-tested starter architecture. I make every architectural, security, and UX decision. Speed without slop.
Production deploy on Supabase + Vercel, your repo, your data, full handover. Optional care plan covers hosting, monitoring, and a feature iteration every month.
Tell me what your business needs — a retention app, an AI agent, a platform you're renting that you'd rather own. I'll send back a free build audit: what it takes, what it costs, and how fast it ships.
// Replies within 24h · Fixed-price scopes · 2 build slots per month
// Full ownership alternative to Skool, Circle & Kajabi
Skool costs $99/mo plus 2.9% of every transaction — forever. Your community lives on someone else's platform, with their branding rules, their feature roadmap, and their right to change pricing. For a coaching business doing $20k+/mo in memberships, that's thousands per year in rent for software you'll never own.
A complete, production-grade community and course platform: email/password and magic-link auth with session persistence, role-based access for admins and members, course and lesson management with drip scheduling, file uploads, member progress tracking, learner submissions with feedback loops, threaded comments, and an admin dashboard with analytics. Modular extensions already built: custom branding, group chat, post scheduling, quizzes & certificates, rich-text lesson completion, RSVP + Stripe payments, and a services marketplace with tracked client-delivery pipelines.
React 18 · TypeScript · Vite · Tailwind + shadcn/ui · Supabase (Postgres, Auth, Storage) · Row Level Security for role-aware access · Vercel
What this means for you: instead of renting your community, you own the entire platform — deployed under your brand, your domain, your data. One fixed price. No per-member fees, ever.
// ghl-api.lovable.app — try it right now
GoHighLevel's API spans 413+ endpoints across 39 domains — CRM, messaging, payments, calendars, and more. Builders waste hours bouncing between docs, Postman, and code editors, guessing at OAuth scopes and required parameters.
A full-stack interactive explorer: browse every endpoint, fill form-generated requests, fire real API calls through a rate-limited proxy with auth validation, and get help from an integrated AI assistant that understands all 413+ endpoints. One-click exports to cURL, Make.com blueprints, and n8n workflows. Multi-provider AI routing across Claude, GPT-4o, Gemini and six more.
React 18 · TypeScript 5.8 · Vite + SWC · React Query · React Hook Form + Zod · Supabase Edge Functions · 9 AI providers
What this means for you: if your business runs on GHL, I know its API deeper than almost anyone building on it. Your app will talk to your CRM natively — contacts, calendars, payments, conversations — not through duct-taped Zapier chains.
// ~80% of Closebot's capability, fully owned
Businesses lose leads because nobody answers fast enough. A receptionist is expensive; generic chatbots feel robotic and can't answer business-specific questions. Tools like Closebot solve this — at a recurring SaaS price, with your conversations on their infrastructure.
A multi-tenant AI front-desk platform where each business gets its own agent with a custom persona and knowledge base. Inbound SMS and GHL messages flow through a multi-stage pipeline: intent classification → RAG knowledge retrieval → self-correction → response generation. It books, reschedules, and cancels appointments synced with GHL calendars. The standout: a self-improvement loop — responses are auto-evaluated, knowledge gaps detected, and client-specific rules evolve weekly without manual tuning.
Claude Haiku (classification) + Sonnet (responses) · RAG with embedded knowledge base · GHL conversations & calendar sync · React admin dashboard · Supabase
What this means for you: an AI receptionist tuned to your scripts and offers, that gets smarter every week — owned outright instead of rented per month.
// The system behind every build on this page
Build AI School is my education platform — marketing site, website-builder wizard, and internal proposal tooling, all running on the same stack I ship for clients. But the real asset is the App Builder: a delivery engine that turns an idea into a production app through a disciplined pipeline.
One planning session produces a complete written plan — data model, screens, integrations, edge cases. Then one feature per session, one PR per feature, with review gates at every step. AI agents write most of the code on a battle-tested starter architecture; I make every architectural, security, and UX decision. If your app has ten features, it ships in ten controlled increments — not one risky big bang.
React 18 + TypeScript starter base · Supabase Postgres/Auth/Edge Functions · Vitest test gates · Claude Code + Codex agent workflows · Vercel
What this means for you: this engine is why I can quote fixed prices and 2–4 week timelines with a straight face. You're not betting on heroics — you're buying a repeatable system.
// The front desk that never sleeps
Missed calls are missed revenue. After-hours callers don't leave voicemails — they call your competitor. Hiring round-the-clock phone coverage costs more than most small businesses can justify.
AI voice agents that answer, qualify, and route live phone calls with natural conversation — not phone-tree robotics. Calls are recorded and transcribed, outcomes sync to the CRM, and qualified callers get booked or routed to a human. Built on production telephony APIs (Wavix, Aircall) with the same conversation intelligence stack as Smartchat: intent detection, business-specific knowledge, and structured handoff rules.
Telephony API integration · Real-time transcription · LLM conversation engine · Call recording & sentiment · GHL/CRM sync
What this means for you: every call answered in two rings, every conversation logged in your CRM, every qualified lead booked — at a fraction of a receptionist's salary.
// From cold lead to booked call, no human required
Speed-to-lead decides who wins the deal, and most businesses respond in hours, not seconds. Human setters are expensive, inconsistent, and asleep when leads come in at midnight.
An AI setter that works leads over SMS and chat: it answers questions, handles objections in your voice, checks real availability on GHL calendars, and books the appointment — confirmation, reminders, and reschedule handling included. The full appointment lifecycle (book, reschedule, cancel) is synced to GHL, with smart follow-up sequences that cut no-shows.
SMS/chat channels · GHL calendar free-slot checks · Objection-handling conversation design · Follow-up & no-show logic · Full CRM logging
What this means for you: a setter that responds in seconds, books while you sleep, and never asks for commission.
// Production client work — details kept confidential
A coaching-style business needed clients to stay engaged between sessions — and needed coaches to see who was on track and who was about to slip away, before they churned.
A client accountability platform delivered as a paid engagement: client dashboards with structured check-ins, task management, and progress views; a coach-side admin with per-client detail pages; and an AI sentiment-scoring pipeline that reads check-in responses and flags at-risk clients automatically. Coaches log weekly, group-call, and 1-on-1 check-ins, track monthly performance numbers with snapshot and trend views, move clients through cohort stages, and read the whole department while editing only their own clients — the multi-coach model real teams need. Includes secure coach impersonation for support, full auth flows, and a tested codebase (unit tests on every core page) running in production.
React + TypeScript · Supabase (Postgres, Auth, Edge Functions) · AI sentiment scoring on check-ins · Vitest test suite · Vercel + edge deployment
What this means for you: this is the retention machine in action — visible progress for clients, early churn warnings for you. And yes, client confidentiality holds: yours will too.
// Small tool, same production standards
Freelancers and small businesses burn time fighting Word templates and clunky invoicing SaaS for something that should take thirty seconds.
A clean invoicing app: manage clients and line items, generate branded PDF invoices in seconds, store and retrieve them securely. Email/password and social auth, type-safe database operations, and row-level security so every user only ever sees their own data — the same security posture as the big builds, applied to a small tool.
React + TypeScript · Supabase (Postgres, Auth, Storage, RLS) · PDF generation · Real-time updates
What this means for you: proof that internal tools don't have to be ugly or insecure. The boring parts of your business are usually the cheapest wins.
// Turn an AI subscription into an operating system
Most teams buy Claude or ChatGPT seats and get a fancy autocomplete. The gap between "we pay for AI" and "AI runs parts of our business" is setup nobody on the team has time to learn: connectors, skills, agents, and workflows.
Full Claude implementations, including my own: custom MCP servers (I've built and shipped several, including a GHL toolkit MCP that gives agents direct CRM control), a library of 60+ custom skills encoding business processes, scheduled AI employees that run recurring jobs autonomously, and Cowork rollouts that connect Claude to Slack, Notion, email, calendars, and databases. This isn't prompt training — it's infrastructure.
MCP server development · Custom skill authoring · Agent orchestration & scheduled tasks · Connector setup (Slack, Notion, GHL, Supabase…) · Team onboarding
What this means for you: the same AI delivery system that lets one builder ship like a team — installed inside your business, with your tools, run by your people.