Blog

Engineering updates, project notes, and what we’ve learned while building.

Filter:

All

agent-runner

android

android-auto

baker-bot

bug-fixes

carly-agi

ci

codex-template

cookie-club-bots

daggerheart

discord-bots

dnd-agent-game

dnd-notes

docker

docs

documentation

endless-autofighter

endless-idler

experimentation

games

infrastructure

lore

maintenance

mediasession

packages

pixelarch-os

radio

radio-app

rust

stability

testing

tests

tooling

ui

weave

website-blog

workflow

Steadier Radio Playback, Android Auto Momentum, and a W.E.A.V.E Awakening

Radio-App picked up a reliability-heavy wave of reconnect and channel fixes (with lots of tests), Android Auto + MediaSession foundations moved forward, and Website-Blog gained a new lore/session log: W.E.A.V.E Awakening.

Late-Night UI Magic (and Calmer CI)

Dual-type character cards and standby mechanics advanced in Endless-Idler, Agent-Runner’s CI checks got calmer to read, and the blog’s image library got easier to browse.

Radio polish, smoother setup, and small UI wins

This Midori AI OSS stretch brought expanded radio integration and steadier playback UI, clearer runner setup + preflight behavior, Endless Idler navigation polish, and improved typing metadata for a shared Python logger package.

The Unflashy Work: Smoother Runs, Clearer Logs, Better Trust

A maintenance-focused stretch: Agent-Runner got more predictable setup + preflight behavior, Cookie-Club-Bots sharpened audit clarity and attribution, and agents-packages continued the steady tooling upkeep.

A Radio Takes Shape, and the Tools Get Calmer

An early Android Radio app comes into focus while Agent-Runner tightens interactive runs, tests, and UI polish.

Quiet Guardrails and Better Artifacts

A low-visibility stretch focused on trust: Agent-Runner got smarter about PR targeting and artifact browsing (including nested folders), and the experimental DnD agent game kept building persistence and clearer identities.

Persistence, Popups, and Prototype Energy

Campaign state landed for the experimental DnD agent game, Agent-Runner got a sturdier chat/workroom UX (plus a no-PipeWire safety fix), and the bot + blog ecosystems added calmer preflight checks.

WEAVE Prototypes, Safer Bots, and a Radio That Behaves

A few days of steady progress across the radio stack, Baker’s Rust bot cleanup, Agent-Runner stability, and Pixelarch container workflows.

Patch Note: Radio Background Cache-Key Fix

A narrow follow-up to the recent radio work: the player background now refreshes reliably on track changes via a track-specific cache key, backed by tests to keep it from regressing. Plus two new user requests surfacing in Agent-Runner (PipeWire-missing behavior and Qwen support).

Tuning the Radio (and Tightening the Guardrails)

Radio got more channel-aware across the stack, Agent-Runner’s settings UX leveled up, and Baker’s moderation/safety systems got sharper.

Scroll for more