Nightly Google Drive voice memo transcription with Cartesia

PromptCreate

Every night at 2am, sweep a designated Google Drive folder for new audio files (voice memos, interview recordings, dictated notes) and drop a text transcript next to each one. This is a deterministic code workflow: list, skip-if-sibling-exists, download, transcribe, write, next.

Trigger: cron at 02:00 in the org's timezone. Track a lastRunTimestamp between runs so each sweep only picks up recordings added since the previous run.

Step 1 — find new audio. Call Google Drive's List Files action with a q filter of parents = '<transcripts-inbox folder id>' and mimeType contains 'audio/' and createdTime > <lastRunTimestamp> and trashed = false. Page through the results (nextPageToken) so no file is missed. Request the id, name, and mimeType fields.

Step 2 — skip if already transcribed. For each returned audio file, derive the transcript name as <base name>.txt (strip the audio extension, append .txt). Before doing anything expensive, run a second Google Drive List Files with q = parents = '<same folder id>' and name = '<base name>.txt' and trashed = false. If it returns any hit, skip the file — we never re-transcribe.

Step 3 — download the audio. Call Google Drive's Download File Content action with the audio file's id to pull the raw bytes.

Step 4 — transcribe. Call Cartesia's Transcribe Audio (STT) action, uploading the audio bytes as the file part and setting the model to ink-whisper. Cartesia's Ink Whisper is a batch STT model that handles arbitrary-length audio, so no chunking is needed.

Step 5 — write the transcript. Take the returned transcript text and call Google Drive's Create File From Text action to save it into the same parent folder as the source recording, named <base name>.txt, with mimeType text/plain.

Then move to the next file. After the loop finishes, persist the new lastRunTimestamp so tomorrow's sweep starts where this one ended.

No summarization, no formatting, no judgement — the .txt is the raw transcript verbatim. The sibling-check makes the whole workflow idempotent: safe to re-run any time without duplicate transcripts.

What does this prompt do?

Scans one Google Drive folder each night for new audio (voice memos, interview recordings, dictated notes)
Runs each new recording through Cartesia's Ink Whisper speech-to-text
Saves the transcript as a plain text file right next to the original recording, in the same folder
Skips any recording that already has a matching transcript so nothing gets transcribed twice

What do I need to use this?

A Google Drive account with the folder you want scanned
The Google Drive folder you want treated as the transcripts inbox
A Cartesia account for the transcription

How can I customize it?

Change the schedule (hourly during work hours, weekday mornings, on-demand)
Point at a different Google Drive folder or narrow the sweep to specific audio types
Adjust the transcript file naming so the .txt lands wherever fits your file conventions

FAQs

What kinds of audio files work?

Anything Whisper-family models handle: voice memos from your phone, Zoom or Meet recordings, dictated notes, podcast raws. As long as it lives in the Drive folder and has an audio file type, it gets picked up.

What happens with really long recordings?

Cartesia's Ink Whisper is a batch speech-to-text model built for arbitrary-length audio, so hour-long interviews are fine. Each file is transcribed end-to-end in one pass.

Will it re-transcribe files if I run it again?

No. Before transcribing, it checks whether a text file with the same base name already sits in the folder. If it does, that recording is skipped.

What does the transcript look like?

A plain .txt file named after the audio file (so my-memo.m4a becomes my-memo.txt), saved right next to the original in the same Drive folder. No summarization or cleanup, just the raw transcript.

Does the folder need to be organized a particular way?

No. Just drop audio into it. The workflow uses a Drive query to find every new audio file under that folder, so subfolders and mixed non-audio files are ignored.

Related templates

Ad creative studio with a pre-approved variant library

Generate on-brand ad variants in every format, approve the winners, and keep a ready library you can swap in the day an ad starts to fatigue.

App

Filter every signed Ironclad contract and brief it on demand

Search, filter and sort your whole signed contract repository, then have an assistant read any contract and write a plain English brief back onto the row.

App

Artwork desk for the Notion posts still missing an image

Open one screen each Monday to see which upcoming posts still have no image, generate three on-brand options, and file the one you pick.

App

Webinar promo desk: find Zoom sessions missing artwork

See every upcoming Zoom webinar that is close, under-registered and still has no promo graphics, then build a full image kit in one click.

App

Turn Google Drive scripts into narrated audio, line by line

One desk to pull scripts from Drive, give every line its own voice, redo the one take that missed, and share approved audio in Slack.

App

Work your Shopify catalog until every product has a clip

See which products still have no video, animate their hero photos into short cinematic clips, and compare motion styles side by side before you pick a winner.