Houjun Liu and Leonid Spektor have created a pipeline for taking the raw audio of a spoken interaction and outputting a CHAT transcript. The system works well and reduces transcription time by roughly 75%. It uses the API of the rev-ai.com online ASR service. You would need to sign up and pay for that service. The system is currently only configured to work on MacOS and it requires installation of various utilities through Miniconda and the Montreal Forced Aligner (MFA) syste. Source code and a readme is available here in the batchalilgn repository. This is very much work in progress and we hope to have distributable versions of the pipeline later this year.