The option for importing saves can be found near the option for exporting. Note that in Top 50, the import buttons are on the row below.
.txt instead of .json should work. Note that the save file can be imported into Livecountsedit perfectly fine as a .txt file. Note that save files created after June 4, 2025 use the MIME type text/plain despite being a .json file in order to get around this issue.
The exact details on how each option works are a bit advanced if you don't know coding, but you can use one of the mixerno.space presets in Top 50, SocialBlade, and Akshatmittal counters, and copy over the corresponding settings from one of these counters for other counters (they will receive the preset feature in the future). Make sure you configure the YouTube channel ID, or create new counters with ID field set to the corresponding ID on Top 50. You can convert a YouTube username to an ID using this tool.
If you use the "Force updates" option, the counter will be the real sub count, otherwise, you can make your own estimations for abbreviated YouTube counts. The SocialBlade and Akshatmittal counters come with an "overestimation leeway" option to specify how much to bring the count down by if you overestimate. For example, if your count hits 101M but the API still says 100M, you can use a 10% leeway to drop it down to 100.9M, or a 20% leeway to drop to 100.8M. Using a 0% leeway will drop to 100,999,999.
Also, remember to save your API settings! They don't get saved automatically.
We used to have a button to automatically import them, but we removed it due to copyright concerns. You can find community-made lists like this one or copy from a save file shared in our Discord server.
Make sure you set the fire icon gain type to "Hour", and add them in order from highest to lowest so they take priority in the correct order.
Watch this video by MDM himself for the exact rates:
Mean and standard deviation gains offer an option for a (possibly) more realistic count growth. For example, instead of being confined to a minimum and maximum gain, you add a mean value, and the standard deviation controls the spread. It is a normal, aka Gaussian distribution, so values closer to the mean are more likely.
For example, if your mean is 20 and your standard deviation is 5, most gains should be within 5 (or 1 standard deviation), so 15-25.
As a rule of thumb, around 68% of the time you get a gain within 1 standard deviation (so 15-25 in the example), 95% within 2 standard deviations (10-30), and 99.7% within 3 standard deviations (5-35).
Even if you use min/max gains, the total gain after a long period of time will be approximately normally distributed thanks to the Central Limit Theorem, so we also use this to simulate offline gains and to estimate gains with our gain calculator. Who knew that we would use actual statistics when talking about "YouTube Stats"?