Introduction

User manual (version 1.1)

Hello users,

We are continuously monitoring feedback/bug reports; please use the link on the top right of the app to report any issues that is impacting your user experience.

Please refer to the Guidance sections by an “i” within the app for instructions

User Tips

  • Double clicking a case allows you to Edit a case quickly

  • Cases are locked after the user clicks “Process” the first time; it will only unlock if the case is finished processing. You will receive an email when a case finishes processing — alternatively, refresh the page after a short period and it will be unlocked again.

  • Submitting observations and interviews will cause the Processing phase to take longer

  • Your case management screen should look like the image below — if it doesn’t, zoom out on your browser

  • Italicized text at the bottom of each section in a report is a summary that uses a large-language model to generate, and will vary each time the report is generated. If you see an inaccuracy within these sections, report it to the team our SMEs will review the prompts.

  • If you believe you’ve run into an error or discovered a bug, please do not delete your case. This makes it next to impossible to investigate the issue 😔

  • Remember you can upload more than one singular parent and teacher reports for rating scales. We’ve done quality assurance with 2 parent reports, 3 teacher reports, and 1 self report (if available) per rating scale (Vineland, BASC, Conners, ASRS, BRIEF), but they must be standalone documents.

  • If you Process a case which has unsupported file types, it will remain Open or error out + result in an unusable output.

  • If you are confident you are submitting inputs that should be supported but the case is still erroring after clicking Process (you receive an email that says the case Failed to process), it is possible that OpenAI’s GPT API was down when you initiated the process and caused a fatal error during eval creation. The GPT service is historically available 99.85% of the time, which means on any given day, there may be 3 minutes of downtime that we cannot control. You can try again later (1 hour from initial try, just to be safe) — or you can reach out to our team with a bug report for help and we will try and replicate your issue.

  • Each report is supposed to generate a score table.

    • If you do not see one for a report, submit a bug report to the team and we will investigate your submission.

Known report related issues/room for improvement:

  • The Assessment List on the cover page will be updated to better match common report convention (CONNERS -> Conners 4)

  • Spacing inconsistencies; please send feedback to the team if there is a constant spacing issue with a high volume reports

  • WJ-IV: outputs for the docx version of this report may be inconsistent in such a way that data extraction for it fails.

    • If this happens, try regenerating the document from the provider’s website and directly submitting that to the Lightner app without uploading it in an intermediary host such as Google Drive or OneDrive (download to your device -> upload directly to our website; skip Google Drive, OneDrive)

    • If it still fails, please submit feedback to the team with the caseID for us to investigate.

    • There is no known fix for this, other than we need more samples.

    • To sample with a WJ that we have tested works with our system, please see the Examples page.