Name | Modified | Size | Downloads / Week |
---|---|---|---|
Parent folder | |||
Kemono.Downloader.exe | 2025-08-01 | 68.7 MB | |
Kemono Downloader v6.3.0 source code.tar.gz | 2025-07-31 | 13.2 MB | |
Kemono Downloader v6.3.0 source code.zip | 2025-07-31 | 13.2 MB | |
README.md | 2025-07-31 | 3.4 kB | |
Totals: 4 Items | 95.2 MB | 0 |
-
Feature: Handling of 403 Forbidden Errors
The application is now more resilient to HTTP 403 Forbidden errors.
Problem: 403 errors were treated as permanent failures, causing downloads to be abandoned.
Solution: A retry mechanism was implemented inworkers.py
:- Identifies 403 errors and logs helpful guidance.
- Treats 403 as temporary; retries download automatically.
- Improved request headers to include a
Referer
.
Note: These errors are caused by the server, not the application. They are often triggered by missing cookies, expired sessions, or temporary rate-limiting. Unfortunately, there's not much that can be done on the client side to completely avoid them.
-
Feature: Multipart Download Controls
Previously, the multipart download feature was a simple ON/OFF toggle. It has been significantly upgraded to give users control over its behavior.
Problem: The feature lacked customization, applying to all files and using the main thread count to determine the number of download parts.
Solution: A newMultipart Download Options
dialog was created (MultipartScopeDialog.py
). When a user enables multipart downloads, this dialog appears and allows them to configure:- Scope: Apply multipart downloads to Videos Only, Archives Only, or Both.
- Number of Parts: Users can now set a specific number of download parts per file (default: 4), independent of the main application's thread count.
- Minimum File Size: A threshold in MB (default: 100). Only files larger than this use multipart downloading.
-
Bug Fix: Invalid Characters in File & Folder Names
Post titles with illegal characters (e.g.,?
,:
,*
) could cause a crash.
Problem: Existing name-cleaning functions were not comprehensive.
Solution: A new functionrobust_clean_name
was added toworkers.py
. It uses a regex to remove all invalid characters. All naming logic now uses this function. -
Bug Fix: PDF Creation TypeError
The "Single PDF" feature crashed due to an API change in thefpdf2
library.
Problem: The function used an outdated keyword argumenttext
instead oftxt
.
Solution: Updated allpdf.multi_cell()
calls inSinglePDF.py
to usetxt=
. -
Feature: Conditional Creator Profile Saving
Problem: A.json
profile was created even for one-off post downloads, clutteringappdata/creator_profiles
.
Solution: Logic inmain_window.py
now detects single vs. full-feed downloads. Profiles are only saved for full downloads.