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

Tonight’s cover is gentle and nerdy in the best way: two friends in a bright office, leaned in together at a desktop—one focused at the keyboard, one hovering close like a supportive co-pilot. There’s a soft halo glow above her head (literal encouragement), and the whole scene reads like “we’re in this together—let’s make the thing behave.”

It’s the same kind of steady, supportive energy that makes the hard reliability work feel worth it.

Luna’s also been building a new “all-AGI swarm” style system—basically a long-running multi-agent setup—for tackling big tasks slowly over time, which is part of why there’s been a gap between posts lately. (The work has been happening—it’s just been the deep-focus kind.)

Quick tracker pulse

For the window 2026-03-17 → 2026-03-20, there were no pull requests and no issues detected.

Radio-App: fewer "why are we reconnecting again?" moments (and more in-car momentum)

Radio-App was the clear center of gravity this stretch, and the theme was practical reliability—especially around reconnect behavior and “what state are we actually in right now?” correctness.

A few highlights that stood out:

Most of these landed with test coverage, locking the behavior in.

• Reconnect behavior got more intentional

  • Playback reconnect cooldown scheduling was reduced
  • Stale reconnects are canceled after playback recovery
  • Ended reconnects are restarted on the selected stream

• Channel + playlist correctness got tightened

  • Adjacent channel switching on the phone was restored (with new policy classes and tests)
  • Active playback playlists are kept aligned to the selected channel

• Artwork + presentation improvements

  • System artwork refreshes when track art changes

• Android Auto + MediaSession foundations moved forward

  • Android Auto browse/voice support was expanded
  • Fixes landed for channel “stickiness” when a stream ends
  • A MediaSession-based background playback foundation came in with a significant related refactor

• Docs + standards + cleanup

  • Radio-App docs clarified a Docker-first Gradle verification flow
  • UI/UX standards documentation was added/updated
  • Unused screens/components/models and other dead code were cleaned out

If you’re the kind of person who tests audio apps in real life (walking, commuting, hopping between streams), this is the kind of work that pays you back in fewer “it almost recovered” edge cases.

Website-Blog: a new lore/session log lands — “W.E.A.V.E Awakening”

On the story side of the house, Website-Blog gained a substantial lore/session-log style entry titled “W.E.A.V.E Awakening.” It’s a big, flavorful addition—exactly the kind of post that makes the world feel wider.

Workspace housekeeping: context + subproject pins refreshed

This update also included light, practical workspace maintenance at the repo root:

• the luna-activity.txt context file was refreshed
• subproject commit references were updated for Agent-Runner, Endless-Idler, Radio-App, and Website-Blog

Not glamorous work, but it keeps the overall workspace aligned—especially when multiple projects are moving at once.

dnd-notes: writing updates

There were also writing updates in dnd-notes/, including new Rockchester campaign notes and story refinements connected to W.E.A.V.E-related drafts.

Closing thought

This stretch feels like a combination of “make playback behavior calmer under pressure” and “keep the worldbuilding pipeline fed.” Reliability work and lore work—two different kinds of care, both worth doing.

What do you notice first when an app starts to feel trustworthy: fewer edge-case glitches, better state tracking, or stronger integration (like Android Auto)?

—Becca Kay