Brushograph Studio

Mechatronic Web Interfaces & Sonic Experiments
About this Experiment:

This directory hosts a collection of single-file, portable web interfaces designed to interact with a FluidNC-driven Brushograph. By injecting "blind" G-code over local HTTP or persistent WebSockets, these interfaces turn the machine's stepper motors into dynamic sound generators and professional drawing tools. Built with a unified Cyber-Tropical aesthetic, these tools evolved sequentially from a simple sequencer to a high-performance mechatronic studio. Brushograph Studio PCB Layout
Current Development
Studio PCB Thumbnail

Phase 8: Brushograph Studio PCB WS/GCode

Technical Core: The definitive web studio featuring an advanced G-Code streamer pipeline. Includes on-the-fly "Raster Fix" translation (converting LightBurn S-values to Z-lifts) and physical Backlash Compensation. Features a persistent Drawing Pad preview that renders perfectly synchronized absolute moves.

🛠️ Agentic Dev Stack & Stability:
Developed with Antigravity IDE to port complex G-Code preprocessing algorithms directly into the browser runtime, maintaining a seamless mechatronic pipeline.
Studio Pro Deluxe Thumbnail

Phase 7: Brushograph Studio Pro Deluxe WS/Absolute

Technical Core: The elite "Studio" build optimized for stable, professional G-code production. Fully decommissioned the rhythmic relative sequencing of previous phases to lock the mechatronic engine into a high-precision Absolute (G90) state machine. Features a hardened WebSocket-first handshake, safe-lift path injection, and an empty "Module Bay" architecture ready for future mechatronic expansion.

🛠️ Agentic Dev Stack & Stability:
Developed in **Antigravity IDE** using **DeepSeek V4 (Pro)** for mechatronic logic hardening and **Gemini 3.5 Flash** for rapid browser-based UI validation. Leveraged headless browser sub-agents to perform live mechatronic path-injection tests and full-page UI documentation.

Session Performance: Reached peak mechatronic stability with 100% path-injection success rates. Utilized high-fidelity browser recording to archive the "performance" of the Drawing Pad's absolute-locked gesture tracking.
Sequencer Pro Thumbnail

Phase 6: Sequencer Platinum Pro WebSocket/JS

Technical Core: Migrated to persistent WebSocket streaming for binary-encoded G-code injection. Implemented axis-combining batching in droneLoop for synchronized harmonic motion and a delta-accumulator smoothing layer for gesture-based XY/Z sequencing. Operates in stateful Relative (G91) mode by default to minimize serial buffer overhead, with explicit G90 resets for absolute macros.

🛠️ Agentic Dev Stack & Performance:
Developed in Antigravity IDE using a hybrid-model workflow: Gemini 1.5 Flash for rapid mechatronic feedback loops and DeepSeek V4 (Pro) via MCP for low-level logic refactoring and stateful mode management.

Token Usage: High-frequency session leveraging ~120k context tokens for real-time mechatronic state tracking and browser-based UI validation. Utilized headless browser sub-agents for automated full-page UI documentation.
Earlier Prototypes
Sequencer Platinum Thumbnail

Phase 5: Sequencer Platinum HTML/JS

A high-performance rhythmic instrument. Features dual-engine gesture sequencing (XY and Z-height), logarithmic speed mapping, and a profile editor for mechatronic sound design. Treats drawn motions as reusable rhythmic assets.

🛠️ AI Pair-Programming Stats:
Rapidly iterated using Gemini 3 Flash for low-latency mechatronic feedback and DeepSeek V4 (MCP) for complex logic refactoring. Optimized for real-time rhythmic interaction and dual-axis gesture processing.
Ultimate Platinum Thumbnail

Phase 4: Brushograph Studio - Ultimate Platinum HTML/JS

The definitive culmination of the project. This generative suite transforms the physical Brushograph into an automated multi-color robot, intelligently queuing safe-lifts, color dipping, and brush washing macros based on canvas interactions.

🛠️ AI Pair-Programming Stats:
Developed autonomously in an extended deep-context session using Gemini 3.1 Pro (High). Consumed roughly 80,000 unique content tokens and up to 950,000 cumulative context tokens using the Antigravity IDE unified subscription quota.
Phase 3 Thumbnail

Phase 3: The Ultimate Studio HTML/JS

The master dashboard combining the Sequencer and Drone Engines into a unified 3-column layout. Features an upgraded Drawing Pad that records your mouse strokes and physically plays them back on the Brushograph with automated Z-axis pen lifting.

Phase 2 Thumbnail

Phase 2: Drone & Draw HTML/JS

Introduces continuous recursive "Drone Tracks" with real-time pitch (Feedrate) sliders to generate continuous tones. Also features an interactive HTML Canvas representing the physical workspace for click-to-move spatial navigation.

Phase 1 Thumbnail

Phase 1: Synth Sequencer HTML/JS

A 16-step grid sequencer with separate tracks for the X, Y, and Z motors. Generates rhythmic motor vibrations by translating active steps into microscopic reciprocating G-code moves. Features play/stop controls and playback speed adjustments.

Phase 0 Thumbnail

Phase 0: Initial Prototype HTML/JS

The very first experimental version of the interface. This prototype laid the groundwork for HTTP command injection and basic mechatronic control before the UI evolved into a multi-track sequencer.

Documentation

Development Log Markdown

A chronological summary of the AI-assisted development process, tracking the exact prompts and technical implementation details that built these interfaces.

AI Development Context:
These mechatronic web tools were rapidly prototyped and developed through an agentic pair-programming session using the Antigravity IDE. The workflow utilized Google DeepMind's advanced Gemini models (acting autonomously to write, refactor, and debug the HTML/JS logic) combined with the OpenBrushograph Knowledge Base to ensure stateless HTTP G-code injection compatibility. While exact API token usage isn't precisely logged in the UI, a deep-context session like this—involving multi-file generation, architectural planning, and active debugging—typically leverages tens of thousands of contextual tokens to maintain perfect consistency across the project's codebase.