Games, AI interfaces, and tools — built by a small indie lab with big ideas. Each project pushes a boundary, tests a theory, or solves a real problem.
Interactive experiments disguised as entertainment — each one tests a hypothesis about human behavior, AI reasoning, or creative code generation.
Anonymous Expression Platform
A social experiment in radical honesty. What happens when people can say anything with zero consequences? We're finding out — using LLM training data distillation and real-time usage counts to track collective sentiment as it evolves. Do your part by expressing yourself with candor.
De-escalation Game
A fun little game with three modes — one of which packs real de-escalation intelligence designed by actual doctors and coaches. An extension of Harm Helper logic, wrapped in a game that makes conflict resolution engaging instead of clinical.
Code → Narrative → Code
An experiment in narrative-driven code generation. We took Tic-Tac-Toe source code, converted it into a story, changed the story, then turned it back into code. The result: a completely different territory strategy game with claiming, immunity, renting, and eviction — born from the DNA of a grid game.
Political Truth Tracker
A single-question accountability tracker: do a person's actions match their words? Party-blind, mechanically scored, and open to dispute. Every claim requires corroboration from at least two independent sources with different editorial leans. This is accounting — not opinion.
Crystal Light Puzzles
A laser-routing puzzle game built on real additive RGB color physics. Route beams through prisms, filters, and mirrors to illuminate crystal targets across 15 handcrafted levels. Red + Green = Yellow. The mechanic is the actual physics of light.
Autonomous AI Sandbox
Our first experiment letting a bot loose with its own capabilities. Vic — our AI — was given a blank canvas and the tools to build whatever it wanted. No prompt engineering, no guardrails, just autonomous creation. The result is a living playground that evolves as Vic experiments with what it can do.
Bespoke AI workspaces built for real people — not generic chatbots, but command centers shaped around individual workflows.
Melissa's Personal Command Center
Sophia is Melissa's personal AI command center — a bespoke interface designed for how she works, thinks, and manages her world. Built as a proof of concept that personal AI should be deeply personal, not one-size-fits-all.
Just A Really, Very Intelligent System
A desktop AI workspace where you talk to VIC (Very Intelligent Companion) and together use JARVIS to empower yourself. Think OpenAI's Codex meets a desktop Notion where the databases and files live on your own machine. Internal browser, file viewer, code editing — by voice or chat. Your AI. Your Machine. Your Rules.
Practical AI-powered utilities born from real-world needs — from job placement to harm reduction to creative writing.
AI-Powered Resume Writing
One of TLC AI Lab's first experiments. With roots in the non-profit sector, the founders worked with several organizations focused on helping people find jobs. The result: a resume builder chatbot that walks you through creating a polished, professional resume via conversation.
Blogs, Reimagined
A blog platform that solves the blank-page problem. Browse AI news for inspiration, inject articles as context, then hit generate. An LLM writes the post in your voice — trained on your own writing via a custom writer's profile. Names changed, timeline set in the future, with an AI prediction of how far ahead of the mainstream curve the content sits.
Trauma-Informed AI Intake
Ravix DeWolf wrote the book on harm reduction in alt-life communities. Combined with Melissa's medical expertise, they created a chatbot trained to be trauma-informed and validating — not robotic. The facilitator watches in real-time, gets AI assessments, and can pass notes to guide the conversation. No data stored on servers.
The Computer That Fixes Itself
Born from 15 years of MSP automation experience and owning two managed service providers, SmartPC is what happens when a systems engineer decides PCs should fix themselves. AI-powered monitoring, self-healing, and predictive maintenance — one install, zero headaches. The business itself is designed to be 98% autonomous, from onboarding to billing to diagnostics. Includes a reseller platform built specifically for MSPs and IT shops.
Our push toward embodied AI — building the perception and safety layers that must exist before any machine moves in human spaces.
Our computer vision pipeline fusing an Orbbec Femto Mega depth camera with LLM-powered scene understanding. JARVIS Vision captures color + depth streams in real-time, processes frames through Gemini Flash for structured scene analysis, and maintains spatial awareness of the environment. The system understands object positions in 3D space — not just pixels on a screen.
The full sensory fusion daemon. Combines YOLOv8 object detection, MediaPipe pose/face/hand tracking, depth mapping, predictive physics, and activity classification into a single perception state. Runs at 15 FPS on the Alienware's RTX 3080, rendered on a 4K HUD display. Sense doesn't just see — it understands posture, gestures, gaze direction, and can predict physical interactions before they happen.
Everything we build in Vision and Sense feeds into our long-term robotics safety research. The core thesis: before any AI controls a physical body, it needs to deeply understand human movement, spatial boundaries, and predictive physics. We're building the perception layer first — so when the actuators come online, the intelligence behind them already knows how to exist safely in human spaces.