Description Symptoms
Fill in Title, Author, and other metadata in the "Book Metadata" section of the wizard
Click "Continue" or navigate away
Values disappear and are replaced with an alphanumeric extraction ID (e.g., 423d828962c34d2b8a53bbe91176305a)
The title intro (if enabled) speaks this hash as individual letters instead of the title you entered
Happens on every step navigation and job creation
Root Cause
build_pending_job_from_extraction() in form.py extracts metadata from the uploaded file
If no metadata is embedded in the file, it falls back to the filename/extraction ID
This extraction-based metadata is set after apply_book_step_form() tries to save user input
User-entered values are overwritten by the extraction fallback
Expected Behavior
User-entered metadata should take precedence over extracted or auto-generated fallbacks
Metadata should persist throughout the wizard and into the final job
Job details should always show the user's title/author, not the extraction hash
Impact
CRITICAL – Users cannot customize audiobook metadata
No workaround (metadata is wiped on every interaction)
Output files have meaningless titles (hash-based)
Title intro speaks gibberish instead of book name
Fix Required
Apply user form metadata after extraction, not before
Use extracted metadata only as a fallback if the user didn't enter anything
Preserve user-entered fields through all wizard steps and job creation
Reactions are currently unavailable
You can’t perform that action at this time.
Symptoms
423d828962c34d2b8a53bbe91176305a)Root Cause
build_pending_job_from_extraction()inform.pyextracts metadata from the uploaded fileapply_book_step_form()tries to save user inputExpected Behavior
Impact
Fix Required