Audio work that stays on your machine
Audio is personal in a way people underestimate. A voice memo is someone’s voice; an interview recording carries a source who was promised confidentiality; a therapy session, a medical note, a legal deposition, a song you have not released — these are exactly the files that free online converters ask you to upload to a server you know nothing about.
Whiztools runs the whole job in your browser instead. A complete build of FFmpeg, compiled to WebAssembly, loads into the page and does the decoding, filtering and encoding on your own processor. Nothing is uploaded, there is no size or length limit beyond your device’s memory, no account, no watermark, and once the page has loaded it keeps working with the network switched off.
Which audio tool you need
- A file someone cannot open: Convert Audio — MP3 for anyone, M4A for Apple, WAV or FLAC for editing
- A file too big to email or upload: Compress Audio, which can aim at an exact megabyte limit
- A recording that starts late or rambles: Audio Cutter to keep one section, Remove Silence to cut every pause
- Levels all over the place: Normalize Audio to a broadcast loudness such as −14 LUFS
- Hiss, hum, fans or traffic: Remove Background Noise
- Several clips that should be one file: Merge Audio, with loudness matched across them
- A lecture or interview to get through: Change Audio Speed, with voices kept at a natural pitch
- Words rather than sound: Audio Transcribe and Speech to Text, both on-device
Formats, bitrates and what actually matters
MP3 is the universal format — nothing refuses it. M4A (AAC) sounds better at the same bitrate and is what Apple devices record. WAV stores every sample uncompressed, which is what editors and transcription engines want and why a ten-minute recording can reach 100 MB. FLAC is lossless like WAV at about half the size, for archiving. OGG Vorbis is the open-format option.
Bitrate matters less than people expect for speech: 64–96 kbps mono is transparent for a voice, and switching a voice recording to mono halves the file for nothing, because a single microphone captured no stereo information anyway. Music is different — 192 kbps is a sensible floor, and 320 kbps is worth it when converting from a lossless source. What you should never do is convert repeatedly between lossy formats: every pass throws away more, so always go back to the highest-quality copy you have.
Loudness, not volume
The most common complaint about home-made audio — "it sounds quiet compared to everything else" — is a loudness problem, not a volume one. Streaming platforms measure a recording’s perceived loudness in LUFS across its whole length and adjust it to their own standard: about −14 LUFS for YouTube and Spotify, −16 for Apple Podcasts, −23 for broadcast. Deliver at the right loudness and the platform leaves your audio alone. Deliver too loud, and it turns you down after your mix has already been squashed to get there.
Browse related how-tos in our Guides, or read more about why Whiztools keeps processing on-device in About.