I got curious about how good text compression algorithms are and found these benchmark results: https://mattmahoney.net/dc/text.html. Note: these are for large files. 

Interestingly [the winner](https://bellard.org/nncp/) ranked by result + decompressor size uses a neural network:

> NNCP is an experiment to build a practical lossless data compressor with neural networks. The latest version uses a Transformer model.

It's also nearly 3x better than gzip.

k00b

aPolitical

This encodes and compresses your message so you can send a longer text over OP_return's 80 bytes limit.

This is an early alpha release for the concept.

https://allesvoorbitcoin.be/toolsguides/MCOR/

https://github.com/avbpodcast/MCOR/releases/tag/1.0.1

bitcoin

This encodes and compresses your message so you can send a longer text over OP_return's 80 bytes limit.
This is an early alpha release for the concept.
made by 'AVB' (Always Value Bitcoin)

alpha version:
https://allesvoorbitcoin.be/toolsguides/MCOR/


github + codebook :
https://github.com/avbpodcast/MCOR/releases/tag/1.0.1