FFMPEG is a program that supports hundreds of Codecs. It's likely they used a H264 or a compatible variant as the actual codec because a) its currently the most commonly used advanced video compression algorithm, b) there is plentiful hardware and software support for encoding / decoding, and c) it has a very good tradeoff for quality, bitrate, and processing horsepower required.
There is no guarantee you will get keyframes using H264 if you are compressing video without detectable screen cuts. Some H264 compliant codecs, like the very commonly used x264 library (used by FFMPEG), do not even need to have dedicated keyframes at all but rather use a technique called Periodic Intra Refresh
to encode videos without keyframes. Periodic Intra Refresh provides much better streaming
/live capture performance since it lowers both encode and decode latency for transmission and it lessens the incidence and severity of data rate spikes when using variable bit rate compression schemes.