Batch transcription: transcribe many recordings at once

Batch transcription, also called bulk transcription, means one setup for many recordings: you transcribe multiple files in one queue instead of one upload at a time. Hushscript stages the files you add, shows what the queue will cost and waits for you to start. It is built for people who already have a stack of files and want transcripts, not for a live meeting feed.

Batch transcription: transcribe multiple files in one queue

The composer holds a queue of staged files. Each file is read first, so the queue knows its length and shows a total estimate. You start the queue when you are ready, or start single files earlier.

Some files are really one recording. A phone recorder that saves in chunks, or an interview split by a restart, can be merged in a chosen order and transcribed as one. The merge is deliberate, never automatic. To work through a whole folder of recordings, select its files together in the file picker. For that workflow in more detail, see transcribing a folder of recordings.

Shared options and per-file starts

The options panel next to the queue applies to every file still waiting. It covers the source language or automatic detection, speaker detection, a saved dictionary, a saved prompt preset, one-off keyterms, multichannel handling and translation targets. Change it and the waiting files follow. A file you have started keeps the options it started with.

Batch tags are read when you add a file, so enter them before you add the files; they attach to each transcript when its job finishes. With speaker detection on, every transcript gets speaker labels you can rename and correct afterward.

File limits for a batch

Each recording can run up to 10 hours. There is no fixed file-size limit, but a large video is prepared in your browser, so the work depends on your device. Video audio is extracted on your device and only the prepared audio is uploaded. Service safeguards also limit how many jobs run at once and how much audio is processed per day, so an unusually heavy batch can queue or need to be started in stages.

What a batch costs

You pay for audio length. Twelve 50-minute interviews add up to 600 minutes, and a 900-minute pack covers them with 300 minutes left. Packs run from $1.99 for 45 minutes to $49.99 for 6,000 minutes, with no subscription. Minutes do not expire, and every file in the batch draws from the same balance at the same rate. If the balance covers only part of the queue, start what fits and top up for the rest. The pricing page lists every pack and pay-as-you-go transcription shows the per-minute math.

Exports for each transcript

Every transcript exports on its own: TXT, SRT, VTT, DOCX, PDF, CSV, JSON and 14 more formats, with presets, an all-files ZIP and a password-protected archive. Tags, search and filters help you pull a set of transcripts together in your list. The export workspace is per transcript, so plan on exporting each one.

Privacy for a whole batch

The same rules apply to every file. Hushscript removes the transcription audio after the transcript is saved, never trains models on your recordings and encrypts stored transcript content. Set auto-delete after 7, 30, 90 or 365 days, or use private mode when no transcript should be saved to the account. Private transcription explains the details and the limits of private mode.

Batches from code

Hushscript has a public REST API and an account MCP server. Both let a script or an agent submit prepared audio and follow each job. Start with the developer docs, or read how automation fits together.

Who uses batch transcription

Batches suit any set of recordings that belongs together: interviews and focus groups for a study, a podcast back catalog, a lecture series, or a season of recorded calls. These pages cover each case:

• Research transcription for interview sets and focus groups.
• Podcast transcription for back catalogs and show notes.
• Interview transcription for a series of conversations.
• Lecture transcription for a semester of recorded classes.

Updated: October 20, 2026