You can definitely do parallel / concurrent / async programming directly in the shell, without any helper programs. It's not exactly ergonomic, but with a few p

You can definitely do parallel / concurrent / async programming directly in the shell, without any helper programs. It's not exactly ergonomic, but with a few patterns it's not too difficult. Async: long_running_cmd & > results.txt cmd_pid=$! # other stuff wait $cmd_pid # do something with results, if you want Parallel: for script in "${scripts[@]}" ; do "$script" & done wait Dataflow: mkfifo my_channel mkfifo another_channel my_cmd > my_channel & tee my_channel \ >(cmd1) \ >(cmd2 > another_channel) & another_cmd < another_channel But if you want a useful dataflow environment, I really recommend checking out [Nextflow](https://www.nextflow.io/ https://www.nextflow.io/). I use it at work for constructing bioinformatics pipelines, and it's really natural. The "preview" DSL2 is worth looking at, as well.