About
A converter that stays on your device.
Flex Convert turns video, photos and audio into formats that open where you need them. The work happens in your own browser, so nothing is uploaded and there is no queue.
Made by Pikosiko.
What it is and how it is made.
How it works without a server
Your browser already knows how to decode and encode much of what you drop on it, through the same codecs it uses to play video. Flex Convert uses those where they exist. Where they do not, for AVIF, WebP in Safari, JPEG and FLAC, a small encoder built into the page does the job and loads the first time it is needed.
That is why the same page works in Chrome, Edge and Safari, and why what a browser cannot write, we do not offer.
What it will not do
It does not write MP3. Browsers cannot, and the free encoders carry licence terms that do not suit a commercial product. It cannot read Dolby Digital or DTS sound, because those codecs are patent-encumbered and we do not ship a decoder for them. It does not convert ProRes. When a file hits one of these limits the page says so, in words, before or instead of producing something misleading.
How the pages are written
A conversion page exists only for something the converter really does, run end to end. Each page carries examples we measured: a script opens the real page in Chrome and in WebKit, drops a test file we generated ourselves, saves the result and checks it with ffmpeg, ImageMagick and macOS tools. The table shows the browser, the day and the machine.
We do not use anyone else's photographs, films or music, and we do not publish a number we could not measure. When a measurement shows the converter falling short, for example the cover art that is not carried into an M4A, the page says so.
Browsers we tested
Google Chrome and Playwright's build of WebKit, the engine behind Safari. That build is not Safari, and we say so wherever it matters. Firefox is not tested and we do not promise it.
Who makes it
Flex Convert is made by Pikosiko, which also makes Flex Remote. It is free to use, with no account and no advertising.
The test files.
Every figure on this site was measured on files made for the purpose. This is the full list, in no particular order.
20 s clip, 1080p, H.264
A generated 20-second 1080p clip: a zoom through a fractal with light film grain and a synthesised chord pad as sound. The same H.264 and AAC streams in a MOV.
20 s clip, 1080p
A generated 20-second 1080p clip: a zoom through a fractal with light film grain and a synthesised chord pad as sound. H.264 and AAC in an MP4.
20 s clip, 1080p, HEVC
A generated 20-second 1080p clip: a zoom through a fractal with light film grain and a synthesised chord pad as sound. HEVC and AAC in a MOV, the way an iPhone set to High Efficiency records.
Transparent logo, 2048 by 2048
A drawn logo on a transparent background, 2048 by 2048 pixels.
20 min clip, 240p
A generated 20-minute 240p test pattern with a synthesised chord pad, H.264 and AAC.
5 min clip, 360p
A generated five-minute 360p test pattern full of fine moving detail, with a synthesised chord pad, H.264 and AAC: a hard clip to compress.
10 s clip, AAC sound
A generated 10-second 720p clip in a Matroska file, H.264 video with AAC sound.
10 s clip, AC-3 sound
The same 10-second clip in Matroska with AC-3 (Dolby Digital) sound.
10 s clip, Opus sound
The same 10-second clip in Matroska with Opus sound.
10 s clip, Vorbis sound
The same 10-second clip in Matroska with Vorbis sound.
90 s music
The synthesised 90-second pad as FLAC.
90 s music
The synthesised 90-second pad as a 192 kb/s MP3 (made with LAME on the test machine; Flex Convert only reads MP3).
90 s music
A synthesised 90-second stereo chord pad, 44.1 kHz, 16-bit, written as WAV.
60 s noise, 24-bit 96 kHz
Sixty seconds of pink noise, stereo, 96 kHz, 24-bit: the hardest content for a lossless coder.
60 s noise
Sixty seconds of pink noise, stereo, 44.1 kHz, 16-bit, as FLAC: the hardest content for any audio coder.
Detailed photo, 12 MP
A generated 12-megapixel picture of fractal detail with fine film grain, the busiest kind of content a camera makes. Carries invented camera and GPS tags.
Detailed photo, 12 MP
The detailed picture as HEIC, written by macOS (sips).
Detailed photo, 12 MP
The same detailed picture saved as PNG.
Smooth gradients, 12 MP
A generated 12-megapixel picture: soft colour gradients, no texture. Carries invented camera and GPS tags.
Smooth gradients, 12 MP
The smooth picture as HEIC, written by macOS (sips).
Smooth gradients, 12 MP
The same smooth gradient picture saved as PNG.
10 s clip, no sound
A generated 10-second 1080p clip with no sound track at all.
90 s of near silence
Ninety seconds that are silent except for a 0.4-second tone every ten seconds, 44.1 kHz, 16-bit WAV.
90 s music with tags and cover
The 90-second pad as FLAC with a title, artist and album tag and a JPEG cover picture.
90 s music with tags and cover
The 90-second pad as a 192 kb/s MP3 with ID3 title, artist and album tags and a JPEG cover picture.
App screenshot, 2400 by 1600
A drawn app screen, 2400 by 1600 pixels: flat colour, text and thin lines.