Troubleshooting
Recover from sign-in, upload, permission and job errors.
Sign-in returns to the login page
Use the same identity provider and account as before. Start again from the Lutrio sign-in page rather than reusing an old authorization URL. If the identity provider requires verification, complete it in a normal supported browser.
A project or dataset does not load
Use the page’s retry/refresh control. Check the project and signed-in account. A permission error should be investigated, not worked around by using someone else’s project identifier. If it persists, record the time, URL and exact message.
Upload reaches 100%, but data is missing
Upload and import are separate stages. Check the background task’s final result, imported count and ignored files. Inspect a few samples. Before retrying, confirm whether the first task already created a dataset.
AI assistance is unavailable
Check the model/service availability shown by the platform. A healthy website does not establish that its upstream AI service is ready. Keep manual edits saved, note the error, and contact support if the same failure persists.
Training cannot start
Check dataset publication, compatible model selection, permissions and compute/credit readiness. Resolve the specific failed condition before trying again. Do not paste arbitrary server paths into model fields.
Task submitted, but no result
Find its existing task record and check whether it is queued, running, completed or failed. Avoid duplicate submissions. When an export completes but history has not refreshed, reopen or refresh the history before creating another export.
Contact support
Email [email protected] with the task ID, page URL without temporary authorization parameters, time and error text. A cropped screenshot is useful if it contains no private samples or credentials.