I wrote Tabrelay, a small MIT-licensed tool that lets scripts and AI agents drive the Chrome you already use and are logged in to: https://github.com/aaronsuns/tabrelay

Why I made it: I didn't want a second "automation browser" to log in to again (2FA, new-device checks, sites that treat it like a bot), and I didn't want a debugging port on my everyday browser that every local program can talk to.

How it works: a Chrome extension plus one Python file (standard library only, no npm, no build step). The command sends jobs (open a page, run JavaScript, click, type, upload a file, screenshot) over 127.0.0.1 to the extension, which runs them with the Chrome DevTools Protocol. Jobs run in background windows so it doesn't steal focus. The extension only touches sites on an allow-list that can only be edited on its own options page, and it only obeys jobs carrying a local token. It's a CLI, not an MCP server, so any agent that can run a shell command can use it.

Honest limits: it's a week old, Chrome/Chromium only, loaded as an unpacked extension (not on the Web Store), a page can tell a debugger is attached, and it has no auto-waiting locators like Playwright. For tests and servers, use Playwright or Puppeteer.

AI disclosure: most of the code, tests and docs were written with Claude Code. I designed it, steer it, review what goes in, and use it myself every day.

Guides and comparisons: https://tabrelay.zexa.se/

[–] [S] 2 points 21 hours ago*

Thanks, that's kind, and fair on trust: don't run it exposed on a stranger's word. It's all in the repo to read. On the YouTube part: the downloader is optional, the server works with just your own music folders, and nothing music-related ships in the repo (https://melarka.zexa.se/youtube-to-self-hosted-music says what it does and where the responsibility lies). You're right that it's a risk for the repo if YouTube tooling draws takedowns, and I'll keep that in mind.

  • source
  • parent
  • context
  •  

    I built a music server for my family and put it under AGPL-3.0: Melarka. Go + SQLite in one Docker container, plus native Android (Kotlin) and iPhone apps, all in one public repo: https://github.com/aaronsuns/melarka

    Why it might be of interest here:

    • Everything is in the repo: server, both apps, docs and the build scripts. You can run it, read it and fork it.
    • It reads your own music folders and never rewrites your files or tags. No account at a third party, no tracking; the only call home is an optional, opt-in install ping.
    • An account per family member (own favorites, playlists, personal radio), synced lyrics, an offline cache for walks, and the Android app starts playing when the phone connects to the car.
    • Honest limits: it has its own API, so it is not Subsonic-compatible, and the iPhone app is not in the App Store (Apple's developer fee for something that earns nothing); you install it from a Mac with Xcode.

    AI disclosure: most of the code, tests and docs were written with Claude Code. I designed it, steer it, review what goes in and use it every day on our phones. I know that is a sore point in open source; the repo says so up front, and I'd rather have the argument in the open than hide it.

    Site with a one-minute tour: https://melarka.zexa.se/

     

    The Android app for Melarka, a music server you run at home (one Docker container). I built both because no Subsonic client held up for my family in the background and in the car.

    • native Kotlin app, not a web wrapper: keeps playing with the screen off, lock-screen and car controls, gapless
    • an offline cache of your favorites for walks with no signal
    • starts playing by itself when the phone connects to the car over Bluetooth, with the current lyric line on the car display
    • synced lyrics, an account per family member
    • interface in English, Chinese and Swedish

    It needs your own Melarka server, so it's no use without one. Signed APK on the GitHub releases page; F-Droid is submitted but waiting until the app is a bit older.

    AI disclosure: built with a lot of help from Claude Code (most of the code, tests and docs); I designed it, review it and test it on our phones every day.

     

    I ran Navidrome for a long time, but never found a phone client that held up for my family in the background and in the car. So I wrote a server and its apps together, and it has replaced Spotify for us.

    • one Docker container (Go + SQLite); it reads your music folders and never rewrites your files or tags
    • an account per family member: their own favorites, playlists and personal radio
    • synced lyrics; anyone can nudge lyrics that run early or late
    • native Android (Kotlin) and iPhone apps: background play, lock-screen and car controls, gapless, an offline cache of favorites, and it starts playing when the phone connects to the car
    • optional: add songs from YouTube (yt-dlp) into your own library

    The catch: it has its own API, so it is not Subsonic-compatible and your existing clients won't work with it. It's about a month old; we use it every day at home.

    AI disclosure: built with a lot of help from Claude Code. It wrote most of the code, tests and docs; I designed it, steer it, review what goes in and test it on our phones every day.

    Site with a one-minute tour: https://melarka.zexa.se/