Customizable output filename

[WillScarlettOhara]

Hello,
I am running tests with the same input EPUB and different models, and I am unable to tell which model is currently being used in the logs, nor for the output file. A suffix with the output language and a number is not sufficient. book (French) (2).epub and book (French) (3).epub are not the most useful information.

I would prefer book (French) (model_used)_date.epub

I could use CLI to name like I want with the --output flag but I prefer to use the webUI.

[hydropix]

Shipped in v1.0.21.

The naming convention field already existed but the frontend was only substituting {originalName}, {targetLang} and {ext} — {model} and {sourceLang} were documented in .env.example but never actually replaced, and there was no date placeholder at all.

What changed:

• Added {sourceLang}, {model}, {date} (YYYY-MM-DD), {datetime} (YYYY-MM-DD_HH-MM-SS) to the web UI pattern substitution.
• {model} values are sanitized for filesystem safety, so anthropic/claude-4.5-haiku becomes anthropic_claude-4.5-haiku and qwen3:14b becomes qwen3_14b — no accidental subdirectories, no invalid chars on Windows.
• The filename is now regenerated at translation launch time instead of at upload time. This was the important one for your use case: previously file.outputFilename was frozen when the file was added to the queue, so changing the model dropdown before clicking Translate had no effect on the output name. Now it picks up whatever model is currently selected.

For your workflow, try this pattern:

{originalName} ({targetLang}) ({model})_{date}.{ext}

Hit Save Settings to .env if you want it to stick across restarts. Grab v1.0.21 and let me know if anything is off.

[WillScarlettOhara]

Thanks, it works. I thought of another useful piece of information to add: the completion percentage, or the number of chunks translated out of the total chunks, or even the number of xhtml files translated in an EPUB.
Thank you.