Hi, you are right! I'm doing my best, really.
Well, the project is more than active. There is currently something to play with, but the engine is not really ready to be used by consumers at the moment.
I have working JIT compiler for solid fills and nearly finished JIT compiler for gradients and textures (all supporting multiple rasterizers, including the agg/freetype one). However, I'm not so far with optimizations as I think this should be the last step, so the engine is mostly using GP registers. What I have now is basically a performance baseline I will use to compare with a more advanced implementation - SSE2, AVX2, and hopefully, AVX-512.
But... even without any advanced optimizations the engine beats all the 2d libraries I have benchmarks for - Cairo, Gdi+, and Qt; where Gdi+ is the fastest from these compared, and surprisingly, Qt is the slowest one.
It mostly depends on your project and the features you need. There will be a pre-release send to interested parties before the code goes to github, but I would like to implement some missing features before this happens.