# Three Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/three
> Markdown URL: https://aitinkerers.org/technologies/three.md
> Technology record last updated: 2026-09-18T15:13:53Z
> Generated: 2026-09-22T09:51:49Z

Three.js is the industry-standard JavaScript library for rendering hardware-accelerated 3D graphics in web browsers via WebGL.

Created by Ricardo Cabello (Mr.doob) and maintained by a global contributor base, Three.js abstracts the complexities of WebGL into a manageable scene graph API. It enables developers to deploy immersive 3D environments (like NASA's Eyes on the Solar System) using standard components: cameras, lights, and meshes. The library supports glTF 2.0 for efficient asset loading and integrates with physics engines (Cannon.js) for real-time interaction. By handling the heavy lifting of GLSL shaders and matrix math, it delivers high-performance rendering on any device with a modern browser.

- Official technology site: https://threejs.org
- Public AI Tinkerers demos and talks: 13
- Result page: 1 of 1

## Recent Public Talks and Demos

### [FORGE: Webcam-Native Hand Control for AI-Assisted 3D Design](https://tokyo.aitinkerers.org/talks/rsvp_07ajHMVupG0)

Still a WIP, you can see the current build in the Video Demo URL. I built FORGE, a browser-based 3D design prototype that lets people modify a procedural model with natural language and manipulate it directly using their hands through an ordinary webcam. During the demo, I’ll show the working system: issuing spoken or typed design commands, translating them into validated parameter changes with Claude, and grabbing, rotating, moving, and scaling the model with one- and two-hand gestures. I’ll also open the code and live vision debugger to explain the architecture, MediaPipe landmark pipeline, gesture state machine, frame-age and inference metrics, operation validation, and the separation between vision, interaction logic, and Three.js rendering.

- Event context: AI Tinkerers Tokyo — October 5 Meetup — 2026-10-05 — Tokyo
- Public talk page: https://tokyo.aitinkerers.org/talks/rsvp_07ajHMVupG0

### [Building Product Videos with Code: From React to Launch Video in 5 Minutes](https://zurich.aitinkerers.org/talks/rsvp_dcRr36x9-9s)

A programmatic video production pipeline using Remotion that lets you build product launch videos entirely in code. I'll walk through a live coding session where we build a video from scratch: composing scenes as React components, reusing existing web app UI components directly in video compositions, generating AI narration with ElevenLabs, syncing text transitions to the beat of a backing track, and rendering the final output. The demo shows how your easy it is to create launch videos for new features, based on the code of the feature.

- Event context: AI Tinkerers Zürich: July Build Night with Swisscom Ventures — 2026-07-01 — Zürich
- Public talk page: https://zurich.aitinkerers.org/talks/rsvp_dcRr36x9-9s

### [I Built An Agent To Be Me On LinkedIn (And It Failed Three Times)](https://boston.aitinkerers.org/talks/rsvp_vNaKz6EaPyc)

A multi-source, voice-conditioned content pipeline that produces scheduled LinkedIn and Twitter posts in my voice — used in production for inbound GTM as a solo consultant. Not "write a LinkedIn post about X." A weekly research-aggregation-and-generation loop with a human editorial gate. What runs end-to-end: Four input streams aggregated weekly: Daily tinkering / build-moment capture (via a custom skill that captures structured notes mid-flow) GitHub repo scan across active projects — surfaces interesting commits / changes / experiments Tunable web research on topics I care about (AI/agents, regulated industries, GTM patterns) Ambient activity log from the week Synthesis layer: all four streams are collapsed into a single weekly research document. This is the structured intermediate artifact everything downstream reads from — it's what turns "the agent posts random stuff" into "the agent posts from a coherent week of actual work." Voice-first writing system: custom voice guide + platform-specific generation guides. LinkedIn and Twitter are not "same model, different length" — they're different generation modes, with the voice guide as the shared spine. Critical reviewer: second agent reviews drafts for clarity, claim quality, and voice fidelity. Outputs edits, not rewrites. Typefully as the HITL editorial gate: scheduled posts land in Typefully. I review, edit, delete, or reschedule. Nothing posts without explicit approval. Important architectural decision — reusing an existing tool as the review UI instead of building one. What I'll show live: The weekly research doc (the structured input artifact — the room can read it) A live generation pass: research doc + voice guide → draft LinkedIn post The Typefully review pass — real editorial decisions, not theatre Side-by-side: a LinkedIn post that lands vs. a Twitter post the system breaks on (the format failure mode) The voice guide diffs I made to fix the "sounds like me but unintelligible" failure mode The public earlier-iteration version is open-source as the voice-first-writing plugin in my GitHub — internal current workflow has evolved past it but the public version is a usable starting point.

- Event context: AI Tinkerers Boston: GTM Agentic AI Launch — 2026-06-29 — Boston
- Public talk page: https://boston.aitinkerers.org/talks/rsvp_vNaKz6EaPyc

### [\[UofT\] Print Anything: Building the Pipeline from Natural Language to 3D Object](https://toronto.aitinkerers.org/talks/rsvp_z0pA85I3o0k)

What if you could 3D print something simply by describing it in plain English, just as you ask ChatGPT a question? That is the idea behind our project. Our system fine-tunes a Qwen 3 model with LoRA in Google Colab to generate ImplicitCAD code from natural language prompts. The generated code is then compiled into STL files, which can be used for 3D printing. ImplicitCAD is a code-based CAD tool, similar to OpenSCAD. Many people in the 3D printing community like it because script-based design allows precise and reproducible measurements. However, the downside is that users need to know how to write code. To address this, we built a two-phase training pipeline. In the first phase, the model learns ImplicitCAD syntax using real examples collected and filtered from the Thingiverse dataset on HuggingFace. This helps the model understand how 3D geometry is written in code. In the second phase, we introduce logical and spatial reasoning problems so the model can go beyond simple pattern matching. In addition to the model, we also built a web-based preview environment and a Dockerized version of ImplicitCAD. This removes the usual edit-compile-wait workflow that makes these tools difficult to use. The project is meant to be a starting point that future teams can build on and extend.

- Event context: AI Tinkerers Toronto - March - hosted by Mozilla! — 2026-03-25 — Toronto
- Public talk page: https://toronto.aitinkerers.org/talks/rsvp_z0pA85I3o0k

### [Sleazy Ron 2: Electric Boogaloo](https://montreal.aitinkerers.org/talks/rsvp_DzkCTEyoD7Y)

Dual Pi Agent Runtime Scheduler Last year, I demoed a multi-agent system designed to generate screenplays for a 3D embodied character named Sleazy Ron, with the goal of live streaming on Twitch. While the approach worked, it felt primitive, output was often mechanical, as the LLM driving the screenplay tended to fall into repetitive patterns. The system was also limited to text generation, with little capacity for broader agent behavior. Since then, I’ve completely reworked the runtime architecture into a dual Pi agent system. The new design separates responsibilities between two agents: Worker Agent (Core Pi) — operates within its own workspace and filesystem, handling reasoning, planning, and long-running tasks. It represents the agent’s “inner thoughts,” capable of performing complex actions like web searches, note-taking, and task management. Performer Agent — runs in a continuous loop, translating the worker’s outputs into real-time expression. It generates dialogue, schedules animations, and controls camera movements dynamically via an event bridge. This separation allows the system to think and act independently: the worker focuses on cognition and problem-solving, while the performer handles embodiment and presentation. To support this, I built a UI that lets me steer the stream by queuing tasks for the worker. It also provides live visibility into both agents, so I can observe their internal state and behavior in real time. The agents both run with Pi Harnesses, powered by Codex models.

- Event context: AI Tinkerers Montreal - March Demo Night — 2026-03-24 — Montreal
- Public talk page: https://montreal.aitinkerers.org/talks/rsvp_DzkCTEyoD7Y

### [An RTS-Style Interface for AI Agents and Codebases](https://berlin.aitinkerers.org/talks/rsvp_n4kJ2vGsbwA)

A desktop app that renders your codebase as a 3D city -- files become buildings, directories become districts. It picks up Claude Code, Codex and Opencode telemetries locally and animates agents flying through the scene as they read, write, and edit files. All local-first -- no data leaves your machine.

- Event context: AI Tinkerers Berlin Meetup - March 11, 2026 — 2026-03-11 — Berlin
- Public talk page: https://berlin.aitinkerers.org/talks/rsvp_n4kJ2vGsbwA

### [Un Agente AI que si puede tocar tu terminal](https://bogota.aitinkerers.org/talks/rsvp_qIkbtvBBLeI)

Un proyecto open Source, que combina análisis estático de Código (tree-sitter) visualización 3D interactiva (three.js) y un sistema de agentes AI tool- use real, que te ayuda a explorar las decisiones de arquitectura.

- Event context: AI Tinkerers Bogotá: El Primer Meetup de 2026 — 2026-02-26 — Bogotá
- Public talk page: https://bogota.aitinkerers.org/talks/rsvp_qIkbtvBBLeI

### [如何利用 kimi k2.5 加速自己的自學 (實戰為主)](https://hong-kong.aitinkerers.org/talks/rsvp_ro6N7CukDoE)

Topic: How to use AI to accelerate Code learning and accelerate exploration on code example: learn something new, like webgpu raytrace diamond TSL shader code Clean the vertex + fragment shader in runtime, gather uniforms With 3 areas of code (vertex, fragment and uniforms) Tell kimi 2.5 it is a diamond shader in three.js. Please help convert the code to TSL three.js Correct syntax and minor mistakes Explore the working code Learn deeper about how it works and explore different changes.

- Event context: AI Tinkerers Hong Kong & GBA: Using AI as a Superconductor for Learning — 2026-02-26 — Hong Kong
- Public talk page: https://hong-kong.aitinkerers.org/talks/rsvp_ro6N7CukDoE

### [Radiology Lab - Medical Imaging Segmentation Lab](https://raleigh.aitinkerers.org/talks/rsvp_M1DsMV5sU6A)

An early preview of a web application built to create segmentation and 3D modeling training data directly within research and clinical workflows in Medical Imaging, using SAM3 and Gaussian Splatting APIs. It supports multiple imaging modalities; including Cardiac MRI, Cardiac CT, CCTA, and Brain MRI. This training data can be quickly plugged in to fine-tune or build machine learning models, which can loop back into the web application. Creating an iterative cycle of improving machine learning models, segmentations and 3D models across all modalities in Medical Imaging.

- Event context: AI Tinkerers Raleigh Meetup — February 11, 2026 — 2026-02-11 — Raleigh
- Public talk page: https://raleigh.aitinkerers.org/talks/rsvp_M1DsMV5sU6A

### [Natural Language to Robot Motions](https://seattle.aitinkerers.org/talks/rsvp_xJ2TlLuu5u0)

Zettaware is an agentic robotics platform for robot motion planning.

- Event context: Frontier Builds Demo Night: Experiments at the Edge of AI — 2025-11-13 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_xJ2TlLuu5u0

### [Meet Ava the Amazing Virtual Avatar!](https://austin.aitinkerers.org/talks/rsvp_jt80spO_meY)

I have built an AI from the ground up using my own models/servers. The backend uses agents (mostly ollama) The frontend is a voice activated interactive avatar using threeJS

- Event context: AI Tinkerers Austin Demo Night — November 10, 2025 — 2025-11-10 — Austin
- Public talk page: https://austin.aitinkerers.org/talks/rsvp_jt80spO_meY

### [What Happens When You Let AI Build a Game?](https://toronto.aitinkerers.org/talks/rsvp_D5YVCtG_VDs)

Over a period of 2 weeks, I used AI (mostly Cursor + Claude Sonnet 3.7 + ChatGPT o1 Pro) to build two playable browser-based games with Three.js and Nuxt.js. These weren’t polished, long-term projects, they were experiments in rapid prototyping through “vibe coding,” where I gave AI rough ideas and let it generate both the requirements and most of the implementation. In this 6-minute demo, I’ll show: - The actual games in action - The workflow I used with Cursor (messy prompts, generated code, and debugging side-by-side with AI) - How the AI tackled surprisingly complex problems like 3D scene setup, multiplayer sync, and game state transitions - What worked, what didn’t, and what it felt like to “co-develop” with an LLM I’ll walk through code, hacks, and surprises, all from the perspective of someone trying to build real things, fast, with AI in the driver’s seat.

- Event context: AI Tinkerers Toronto - April 2025 Meetup at Shopify — 2025-04-24 — Toronto
- Public talk page: https://toronto.aitinkerers.org/talks/rsvp_D5YVCtG_VDs

### [How to direct twitch streams with a multi-agent cinema crew](https://montreal.aitinkerers.org/talks/rsvp_8m3Y7xD9Pwc)

I wanted to build a twitch streamer in a 3d environment that I could control with AI. This led me to develop a multi agent system in langgraph consisting of a director, writer, cinematographer, animation tech and audio engineer. All working together to build 1-2 minute screenplays that can be chained.

- Event context: AI Tinkerers Montreal - March 2025 Meetup at Shopify — 2025-03-12 — Montreal
- Public talk page: https://montreal.aitinkerers.org/talks/rsvp_8m3Y7xD9Pwc

## Related Technologies

- [React](https://aitinkerers.org/technologies/react) ([Markdown](https://aitinkerers.org/technologies/react.md)) — 220 public demos
- [Codex](https://aitinkerers.org/technologies/codex) ([Markdown](https://aitinkerers.org/technologies/codex.md)) — 44 public demos
- [Docker](https://aitinkerers.org/technologies/docker) ([Markdown](https://aitinkerers.org/technologies/docker.md)) — 147 public demos
- [GLSL](https://aitinkerers.org/technologies/glsl) ([Markdown](https://aitinkerers.org/technologies/glsl.md)) — 2 public demos
- [GSAP](https://aitinkerers.org/technologies/gsap) ([Markdown](https://aitinkerers.org/technologies/gsap.md)) — 4 public demos
- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 662 public demos
- [Rust](https://aitinkerers.org/technologies/rust) ([Markdown](https://aitinkerers.org/technologies/rust.md)) — 56 public demos
- [Sonnet](https://aitinkerers.org/technologies/sonnet) ([Markdown](https://aitinkerers.org/technologies/sonnet.md)) — 25 public demos
- [SQLite](https://aitinkerers.org/technologies/sqlite) ([Markdown](https://aitinkerers.org/technologies/sqlite.md)) — 28 public demos
- [TypeScript](https://aitinkerers.org/technologies/typescript) ([Markdown](https://aitinkerers.org/technologies/typescript.md)) — 205 public demos
- [Agent](https://aitinkerers.org/technologies/agent) ([Markdown](https://aitinkerers.org/technologies/agent.md)) — 12 public demos
- [Anthropic Claude Opus 5](https://aitinkerers.org/technologies/anthropic-claude-opus-5) ([Markdown](https://aitinkerers.org/technologies/anthropic-claude-opus-5.md)) — 1 public demo
- [Claude](https://aitinkerers.org/technologies/claude) ([Markdown](https://aitinkerers.org/technologies/claude.md)) — 174 public demos
- [Claude-3](https://aitinkerers.org/technologies/claude-3) ([Markdown](https://aitinkerers.org/technologies/claude-3.md)) — 110 public demos
- [Claude (Anthropic)](https://aitinkerers.org/technologies/claude-anthropic) ([Markdown](https://aitinkerers.org/technologies/claude-anthropic.md)) — 2 public demos
- [Claude Code](https://aitinkerers.org/technologies/claude-code) ([Markdown](https://aitinkerers.org/technologies/claude-code.md)) — 215 public demos
- [component](https://aitinkerers.org/technologies/component) ([Markdown](https://aitinkerers.org/technologies/component.md)) — 1 public demo
- [Daily](https://aitinkerers.org/technologies/daily) ([Markdown](https://aitinkerers.org/technologies/daily.md)) — 2 public demos