Introducing audio + whisper nodes

We are happy to announce the arrival of the audio and whisper nodes. The simplest complete use case is transcription as text. In this functional documentation, we invite you to:

• paste your OpenAI API key
• upload an mp3
• click generate report

Chatting with an audio file

As always, the whisper node plugs into other existing nodes. So for example, it's easy to connect its output into the chat node, and chat with the audio file.

• paste your OpenAI API key
• upload an mp3
• click generate report
• ask the chat to summarize the contents

Integration with Video pipelines

It is worth noting that in the context of Talk to the City reports, the sole purpose of the Whisper node is extracting exact transcripts + timestamps for Video report generation. When the response_format is set to custom, you can enter the interviewee name, as well as the vimeo link. The whisper node then outputs CSV data that can be used directly for cluster / argument extraction and report generation.

We invite you to:

• paste your OpenAI API key
• upload an mp3
• enter the interviewee name
• enter the vimeo link
• click generate report

Your CSV data is now ready for report generation.