I recently set up a Plex server with containerised applications — Plex, Sonarr, Radarr, Prowlarr, and Syncthing. A previous attempt months earlier had failed due to outdated guides, mismatched versions, and documentation written for different configurations.
This time, rather than consulting forums and YouTube videos, I used ChatGPT during active troubleshooting. I created a private project, shared screenshots, and asked specific questions as problems arose — permissions errors, folder mappings, and container visibility issues. While ChatGPT gets things wrong, the experience was like having a very capable tech mate sitting next to me while I worked through it, rather than decoding advice written for someone else’s setup.
The work involved 12–14 hour days debugging:
- Wrong container versions requiring rebuilds
- User permission issues
- Complex folder mappings across containers
- Plex seeing files Radarr couldn’t access
- Radarr downloading content Plex ignored
- Syncthing executing commands incorrectly
Rather than simply copying steps, I reasoned through why failures occurred, which meant genuine learning.
The outcome: a functional Plex setup I actually understand. ChatGPT helped document the configuration from our conversation history, creating reference notes for future troubleshooting.
The breakthrough wasn’t technical — it was about solving problems with a tool rather than jumping between disconnected advice sources. ChatGPT didn’t magically resolve issues; it helped me think, test, and keep moving.