← Documentation

PLUXIX GUIDE

Troubleshooting

Use health checks, browser status and local logs to isolate the first failing boundary.

The website does not load

Confirm the public domain, refresh the page and inspect the browser network for a failed HTML, CSS or JavaScript asset. If the page shell loads but content fails, check the API health status and sign in again.

  • Do not use internal ports such as 3000, 3001 or 4100 in production URLs.
  • A 401 usually means an expired or missing session.
  • A 5xx should be correlated with the server health and application log.

The Agent stays on Starting

Check that the installed Windows service is present and running, then reopen the UI. The local runtime log is under `%APPDATA%\PLUXIX Agent\runtime\logs`. A repair should not create a second service or expose a random public port.

  • Close old PLUXIX Agent processes before a repair.
  • Use the latest published build and confirm its checksum.
  • If the service starts but the cloud is unreachable, test account connection separately from local workspace startup.

A task is stuck or interrupted

Inspect the last conversation event and service health before retrying. A retry should be idempotent; confirm whether a file write, command or external action completed before repeating it.

  • Check internet stability and TLS access to pluxix.com.
  • Check the selected mode and available credit.
  • Check the individual MCP server if the failure occurred during a connected-tool action.

What to send support

Send the approximate time, route or task stage, release version and a redacted error message. Do not send passwords, cookies, tokens, API keys, private repository contents or unredacted logs.