# Turn voice memos and audio recordings into an ebook

Upload audio source material to Scribora, let it be transcribed, and use spoken expertise as the basis for a nonfiction book.

Canonical: https://www.scribora.ai/docs/sources/voice-memos
Category: Sources
Last reviewed: 2026-09-29

## Record the knowledge you would otherwise never write down

Audio is useful when explaining a subject out loud is faster than writing it. Scribora's upload flow can process audio into transcript-backed source material, so a recorded explanation can participate in the same knowledge base as notes, links, and documents.

A focused recording produces better source material than an unstructured hour of thinking aloud. Record one problem, framework, case study, or chapter-sized idea at a time when possible.

## Audio processing is asynchronous

Audio transcription can take longer than text or PDF parsing. Wait until the upload queue shows the file as Added before you build or rebuild the knowledge base.

> The recording is an input source. Scribora writes the book from the transcribed material; it is not simply an audio-to-text export tool.

## Supported audio and size behavior

The current upload control accepts MP3, WAV, M4A, OGG, and WebM and enforces a 50 MB file limit. Audio is transcribed with the server's transcription pipeline before it becomes a gathered source.

For audio larger than the transcription provider's per-request limit, the server attempts to split the recording into smaller segments without re-encoding, transcribes them in order, and concatenates the text. If safe splitting is unavailable or the file cannot be read, the upload returns a user-facing processing error.