[QUESTION] Can primer sequences be trimmed directly from raw fastq data?

[molly-hetheringtonrauth]

The current pipeline uses tiled amplicon sequencing data and therefore the primer sequences on the ends of reads must be trimmed, which the current pipeline uses ivar (soft clip) and jvarkit (hard clip) to do. I'm wondering, if to avoid performing an alignment and then converting the bam file back to a fastq file, primer sequences could be trimmed directly from the raw reads using a program like Cutadapt (https://cutadapt.readthedocs.io/en/stable/guide.html#non-internal)? What are the benefits of performing an alignment prior to primer trimming? Thanks in advance for your response!