Converting on your device vs uploading: what we measured
Online converters usually send your files to a server and send the result back. We convert on your device instead. This page shows our measurements, the conditions they were taken in, and what they do and don't prove.
How we measured
The measurements come from the script scripts/bench.ts in the project, run in headless Chromium on an Apple M4 with 10 cores on 23 October 2026. The timer starts when the files are chosen and stops when every result is ready to download. It includes the first download of the codecs from a local server.
Test images were synthetic 12-megapixel photos, 4032 × 3024 pixels.
The results
| Task | Total | Per photo |
|---|---|---|
| HEIC to JPG, 1 photo | 0.8 s | 0.80 s |
| HEIC to JPG, 10 photos | 1.8 s | 0.18 s |
| HEIC to JPG, 50 photos | 7.3 s | 0.15 s |
| Reduce 1 photo to 100 KB | 1.3 s | 1.29 s |
| Compress 10 JPGs | 2.3 s | 0.23 s |
| Resize 10 photos to 1920 px | 1.3 s | 0.13 s |
Notice how the per-photo time drops as the batch grows. The first photo pays for loading the converter, and then several photos are processed at the same time on separate cores.
What uploading adds
An upload-based service has to receive your files first. That does not depend on its servers being fast. It depends on your connection.
Take 50 photos of 1.5 to 3 MB each, which is typical for HEIC files from a 12-megapixel phone. That is 75 to 150 MB. On a connection with 10 Mbit/s of upload speed, sending 75 MB takes about 60 seconds, and 150 MB takes about two minutes, before any conversion starts and before the results come back down. This is arithmetic, not a measurement of a particular website. Faster connections shrink it, and slow mobile connections stretch it.
What these numbers do not show
- They are from one machine in a test environment, not from a field study of real devices.
- They say nothing about the quality of other services’ output.
- On a very old phone, a large batch can take much longer and uses a lot of memory, which is why large batches are processed one after another on such devices.
The other half of the trade
Speed is the lesser reason to convert on your device. The bigger one is that the photo never leaves it: no copy sits on a server, nobody has to be trusted to delete it, and there is no queue. You can read how that works on the About page and check the numbers for yourself in the usage statistics, including the count of files uploaded to a server.
Try it with your own photos
Drop your HEIC photos here
or
You can also paste an image with Ctrl+V.
Converted on this device. Nothing is uploaded.
Frequently asked questions
Will my computer be as fast as yours?
Probably not. The numbers come from a recent laptop with 10 processor cores. Older phones and computers will be slower, but the work scales with their cores.
Why is the first photo slower?
The converter code has to be downloaded and started once. After that, the next photos reuse it.
Are the images real photos?
No. The test images are synthetic 12-megapixel pictures (4032 × 3024), the size of a typical phone photo. Real photos behave similarly but vary with content.
Did you compare with other services?
Not in these numbers. Other services depend on their servers and on your connection, so we explain the arithmetic instead of publishing a single winner.
Learn more
- How to remove the location from photosPhotos can record exactly where they were taken. How to check, how to remove the location on iPhone, Mac and Windows, and how to clean many photos at once.
- What is EXIF data?EXIF is the hidden information stored inside photos: camera, date, settings and sometimes GPS. What is in it, why it exists and when you should remove it.