New tools, improvements and fixes on FreeSoundTools, by date. Korean-language updates are on the Korean changelog.
October 25, 2026
- New · Audio to Text saves recognized sentences on this device as it goes. If the tab closes, drop the same file again to continue where it stopped, or open the saved result.
- New · Model manager: see the AI models downloaded on this device with their sizes, and delete them.
- New · Guides: Studio sound free online, record, clean, boost and export, remove 50/60 Hz hum.
- Improved · Audio to Text, Subtitle Generator and Meeting Notes now run with cross-origin isolation on the English site, which lets the speech model use more threads.
- Improved · While an AI task runs, the browser warns before you close the tab, and the screen stays on where supported.
- Improved · Files with no speech (silence, beeps, steady tones) now show "No speech detected" instead of made-up text.
- Improved · Results from Voice Noise Remover and Studio Sound can be sent straight to Volume Booster, Audio Converter, MP3 Cutter or Audio to Text.
- Fixed · Cutting or converting with sample rate "Original" now keeps the original rate (for example 44100 Hz).
- Changed · MP3 files saved from the High-quality AI voice include an ID3 tag saying they are AI-generated.
- Changed · Duplicate landing pages (for example /en/tools/noise-cancellation/) now redirect to the main tool page.
- Fixed · Broken help and bug-report links, and broken styles and links in three English guides.
October 6, 2026
- New · Recent tools and favorites: star a tool next to its title and it appears at the top of the home page with your recent tools. Stored only on this device.
- New · Some tools remember the last settings you chose.
- New · Upload check badge: tap the "0 file uploads" badge on a tool page to see the requests the page made, by type.
October 5, 2026
- Changed · Every tool that saves MP3 uses our own MP3 encoder.
- Changed · The Terms include the use restrictions for the High-quality AI voice.
- Improved · MP3 Cutter and the audio editor: typed time input, zoom, ruler, keyboard control, fade length and MP3 as the default format.
- Improved · Text to Speech gained the High-quality AI voice. The model downloads once.
October 4, 2026
- Design · New site design.
Found a problem or want a feature? Write to [email protected].