Reading the logs

The agent keeps two logs, and picking the right one is most of the work.

  • Error log — what went wrong. Start here.
  • Request log — what was asked of the agent, and when. This is how you prove a request from the browser actually arrived, which splits "the agent is broken" from "the request never got there".

Both are visible in the configurator, on the Logs tab.

Where the files are

Logs are written next to the agent, in a logs folder, one file per day:

logs/error_log_2026-08-06.json
logs/request_log_2026-08-06.json
  • Windows, installed — inside the install folder, by default under %LOCALAPPDATA%\Filemote.
  • Windows, portable — inside the folder you extracted.
  • Linux, service install — under the install prefix, normally /opt/filemote. The service also logs to journald, which is usually the faster read:
sudo journalctl -u filemote-agent -f
  • macOS — ~/Library/Logs/Filemote/filemote-agent.log for the launchd agent's own output.

Files older than 30 days are deleted automatically, so the folder does not grow without limit and last month's problem is not still there to confuse this month's.

Reading them

Both are plain text, one line per event, each stamped with the time it happened — the .json in the file name is a leftover, not a format, so reach for a text search rather than jq. The two questions worth asking of them:

  • Does the error log show anything at the time the problem happened? An error at the right moment is usually the answer; an empty log at the right moment means the agent never got the request.
  • Does the request log show the request? If it does, the problem is on the agent's side of the line. If it does not, look at the connection first — the device shows as offline.

The AI service logs separately

The AI service keeps its own output, because it is a separate process with a very different kind of failure — models not fitting in memory, dependencies missing. When AI is the problem, that output is the one that matters; the AI Hub surfaces its errors as they happen, and the AI does not load covers what they mean.

What to send when asking for help

Attach:

  • the error log for the day it happened — not the whole folder;
  • the request log for the same day, when it is a "nothing happens" problem rather than a crash;
  • your version, from the configurator;
  • your platform, and whether the agent runs as a service, as a launchd agent or by hand.

What logs do not contain

  • Not your files. Paths and operations, yes; contents, no.
  • Not your password. It is stored through the platform's protection and never written to a log.
  • Not your encryption key. It exists in memory and in the pairing code, and nowhere else.

Paths are still information — a log names folders you have shared. Look before you attach it to a public forum thread.

Updated 2026-08-06

This didn’t help — contact support