My twtxt.txt feed grew over 700 KiB, so I thought I write a Python script to split it into monthly archive feeds. Now the only things left are automating the execution, automatically rewriting the main feed as well and adding support in tt, so that the full feed is reconstructed in the cache.