AMBERSCRIPT TRANSCRIBES INTERVIEWS. A basic requirement for writing a good interview is to transcribe it—to turn the entire conversation we’ve had with a minister, a big-name actor, the league’s top scorer, or a writer into text. Once every word spoken by the interviewee has been written down, it becomes much easier to extract a concise summary. Not only that: the summary will also be more faithful to the interviewee’s actual words and reasoning.

The problem is that transcription takes a long time and clashes with the rhythms of journalism, especially online. That’s where platforms like the one run by the Dutch startup Amberscript B.V. come in, powered by Artificial Intelligence. Essentially, we upload the conversation file to Amberscript’s website—or app—and wait for an accurate transcript, ready in seconds.

Amberscript also distinguishes between different voices in the conversation (for example, the journalist and the interviewee). It does this in 39 languages worldwide, including Italian. Once the transcript is generated, we can place the cursor on any point in the written text and a second marker will automatically jump to the same point in the audio track.

The website and app accept audio files in M4A, MP3, and WAV formats, as well as video files in M4V, MOV, and MP4. Transcripts can be exported in Word, JSON, SRT, VTT, EBU-STL, and plain text. Amberscript’s services can be tested for free.