Tech, dev & geek stuff.

Its defining capabilities include real-time tempo changes without altering pitch, optional varispeed behaviour, automatic BPM detection, and per-loop audio quantisation that time-warps recorded material towards selectable rhythmic grids. A Merge function can seamlessly bounce multiple playing loops into a single slot, supporting evolving arrangements during performance.

The system is designed for up to eight stereo loop slots across two banks, with MIDI clock synchronisation, saved performances, imported audio and flexible stereo or mono routing. Extensive customisation options include assignments for footswitch taps, external controls, LED colours and interactions, display behaviour, and detailed settings for the entire looper engine. A companion desktop and mobile application provides deeper configuration, control, audio loading and preset management.

Scheduled for release in late 2026.

Pending public release. The app will be used in the performance in November 2026.

It provides seven pages of 4x4 drum-pad groups. Pad triggers appear in both the device view and pop-up window, while MIDI-mappable controls allow page changes between Drum Rack slots.

Window placement and open or closed state persist when a session is reloaded, and pad names update automatically to match the samples in the Drum Rack.

Metadata stored in each clip's name is displayed for the selected clip, along with the clips immediately above and below it.

In my setup, it shows the song title, performance notes, original clip BPM and the instrument I should be playing. The displayed information is fully customisable, making it useful for performers who need reminders or contextual information within a live set.

It was built from the voices of professional vocalists, recorded during illness. Not studio-bright voices, not polished choir vowels, not the clean sustained notes people usually try to capture. These were rough voices: broken, tired, airy, cracked, and barely able to sing.

It was created from curiosity, really. I noticed that when a voice is sick, I often like its timbral quality. Notes pull apart. The air changes. Unusual overtones appear, creating a strange and airy richness. Quarantones Choir was made from those moments.

The result is a vocal instrument with a soft, beautiful, but slightly damaged character. I've used this a ton in my own sound design and compositions, and I hope you enjoy it too.

CaveCode is a source-understanding tool for C and C++ codebases. It parses real source files and produces compact, structured summaries of function behaviour: guards, branches, returns, calls, scalar state, field reads and writes, loops, switches, and unresolved control-flow regions.

It is designed for software developers, embedded systems engineers and LLM coding agents. The aim is to help users quickly understand large or unfamiliar C and C++ codebases. CaveCode provides a first-read map of a file or project, showing the important logic, exits, side effects and uncertainty before changes are made.

Additionally, when used with LLMs, CaveCode provides a compact, legible and digestible source-backed understanding layer for C and C++ code. This drastically reduces token counts and context-window blowouts, which are common when reading C code. For example, in an extremely complicated Buffer subsystem from one of my projects, the scan-project --brief command reduced approximately 124,000 estimated raw-source tokens to 8,800 tokens, a reduction of approximately 93%, while still surfacing the highest-risk functions, guards, calls and outcomes.

CaveCode uses a generic Tree-sitter backend and a domain-neutral fact/regime pipeline. This makes it useful both as a standalone developer tool and as infrastructure for AI-assisted code archaeology, refactoring, and safety checks.

CaveCode is released as open source on GitHub, with a permissive licence for both personal and commercial use. CaveCode is distributed as a CLI tool, and its source code and documentation are available for anyone to explore, modify and contribute to. Contributions for additional features, language support or integrations are welcome on the GitHub repository.

Core features:

• Per-function summaries for C/C++ files
• Guard and early-return detection
• Scalar dataflow at return points
• Struct field read/write auditing
• Enum and struct inventory
• Cross-file call indexing
• Dead helper detection
• Format-string audits
• Compact explanation mode for large functions
• JSON output for programmatic use