"If I only had a brain..." - Your subtitle automation, probably
The dashboard that gives Subgen a brain.
You're watching TV. German drama. Intense scene. Emotional dialogue.
There's a break in conversation and your TV starts throwing random crackhead rabble at you
[Subtitle appears during silence]
"Thank you for watching"
"Subscribe to our channel"
"♪ music playing ♪"
"[inaudible]"
"..."
What. The. Fuck.
Meanwhile, when there IS dialogue:
- Japanese subs fly by at light speed (unreadable)
- German formal/informal completely lost (Sie? Du? Who knows!)
- VRAM maxed out (OOM errors)
- Same settings for everything (one size fits none)
Your subtitle setup is a crackhead. Time to evict it.
Take over the subtitle crackhead game.
Subbrainarr doesn't compete with other subtitle tools.
It dominates them.
- ✅ All crackheads evicted (silence detection that works)
- ✅ Pure product only (no hallucinations, no garbage)
- ✅ Rule your empire (multi-language mastery)
- ✅ Serve quality (every language optimized)
You're not managing subtitles anymore. You're running an operation.
Before: Same settings for everything
- Japanese anime gets verbose German-style translation
- German drama gets clipped Japanese-style subs
- Everything is wrong
After: Each language gets optimal settings
- Japanese: Concise subs (readable at normal speed)
- German: Nuanced translation (Sie vs du preserved)
- French: Proper formality levels
- Automatic. Zero manual switching.
Before: Crackhead subtitles during silence
"Thank you for watching"
"..."
"♪ music ♪"
After: Clean silence
[nothing]
[as it should be]
[peaceful]
Potato GPU (GTX 1660, 6GB):
"Use medium model with int8. Large will OOM."
Sweet Spot (RTX 3060, 12GB):
"Perfect for large-v3! Want to try beam_size: 6?"
Beast Mode (RTX 4090, 24GB):
"🦑 RELEASE THE KRAKEN"
- See what's processing (real-time, not "check logs")
- Queue management (drag to reorder, not edit YAML)
- VRAM monitoring (graphs, not nvidia-smi spam)
- Mobile-friendly (tune from toilet, we don't judge)
Wizard: "Are subtitles too fast?"
You: "Yes"
Wizard: "Increasing length_penalty 1.3 → 1.4"
Wizard: "Test on sample file first?"
You: "Sure"
*3 minutes later*
Wizard: "Better?"
You: "Perfect"
Wizard: "Applied. You're welcome."
Docker Compose (because we're not savages):
version: "3.8"
services:
subgen:
image: mccloud/subgen:latest
container_name: subgen
# ... your existing subgen config ...
subbrainarr:
image: coaxk/subbrainarr:latest
container_name: subbrainarr
ports:
- "9008:9001"
environment:
- SUBGEN_URL=http://subgen:9000
volumes:
- ./subbrainarr:/config
depends_on:
- subgendocker-compose up -dOpen http://localhost:9008
That's it. We're Docker Hub only. Deal with it. 🐳
- Connection: Auto-detects Subgen (or enter URL)
- Library Scan: Analyzes what languages you have
- Smart Setup: "Found German, French, Japanese. Apply optimized settings?"
- Done: Start processing
Time to first subtitle: 3 minutes.
Share your tuning expertise. Get credited. Help the family.
Browse and download community-contributed language profiles:
- "German Drama Optimization" by @hansgruber
- "Anime Settings (Crunchyroll Quality)" by @weebmaster3000
- "Parisian French Formal" by @lefrog
Submit your own profiles and get featured in the app.
Works with your stack:
- ✅ Plex / Jellyfin (webhooks, library scans)
- ✅ Bazarr (detects requests, shows source)
- ✅ Sonarr / Radarr (coming soon)
- ✅ Any Subgen instance (we're not picky)
Doesn't break anything:
- Subgen keeps working if you turn off Subbrainarr
- All changes are backed up
- You approve everything
Subgen = Scarecrow (works hard, feels dumb)
"If I only had a brain..."
Subbrainarr = The Wizard (gives it intelligence)
Turns out, Subgen was capable all along. Just needed to unlock it.
v1.0 (Launch)
- 🎯 Per-language optimization (18 languages)
- 🧠 Hardware detection & recommendations
- 📊 Real-time dashboard
- 🎛️ Interactive tuning wizard
- 🔇 Silence hallucination elimination
- 📱 Mobile-responsive
- ⌨️ Keyboard shortcuts (Cmd+K everything)
- 🌙 Dark mode (only mode - we're not animals)
- 💾 Settings backup/rollback
- 📈 Stats & analytics
v1.1 (Soon™)
- 🧪 A/B testing (compare settings)
- 👥 Community profiles (share configs)
- 🔔 Webhooks (Discord, Slack, etc.)
- ⏰ Scheduling (process overnight)
- 🔄 FFsubsync integration
What we collect: Nothing. Zilch. Zero. Nada.
No telemetry. No analytics. No phone-home.
Your data stays on your server where it belongs.
We're not the CIA. We're not the NSA. We're definitely not running a submarine operation out of a private island.
Your subtitles are your business. 🔒
If three-letter agencies want your German subtitle preferences, they'll have to ask nicely (we also accept XRP, no more BTC since we found out Satoshi is actually Jeffrey Epstein).
Created by: @coaxk
Powered by: McCloud/subgen - The excellent subtitle automation that makes this possible. Seriously, go star their repo. Without Subgen, we're nothing.
Built with: React, FastAPI, SQLite, too much coffee, Claude's wisdom, and a healthy disrespect for the status quo
Inspired by: Everyone who's edited compose.yaml at 2am, cursing at subtitle timing
Special thanks: The Wizard of Oz, for the metaphor. The self-hosted community, for existing.
- 📖 Docs: subbrainarr.com/docs
- 💬 Discord: discord.gg/subbrainarr
- 🐛 Issues: GitHub Issues
- ☕ Donate: buymeacoffee.com/coaxk
MIT - Do whatever you want. We're not your dad.
Stop fighting with YAML files at 2am.
Stop googling "whisper hallucination silence fix" for the 47th time.
Stop using the same settings for 12 different languages.
Your subtitles deserve intelligence. Give them a brain.
[Download] [Docs] [Discord] [Star on GitHub]
"There's a break in conversation and your subtitles stop throwing random crackhead rabble at you."
Welcome to Subbrainarr. 🧠