Freya Project — Podcast Transcription & Speech Annotation — OneForma

Last verified: 30 July 2026 · Source: oneforma.com/jobs/freya/

What is this project?

Long-form audio transcription and speech annotation — you're working with 30-to-60-plus-minute podcast recordings, transcribing them accurately and labeling the speech, to help train and evaluate voice recognition and speech-to-text models.

What would I actually do day to day?

• Transcribe long-form podcast recordings following detailed project guidelines.
• Identify different speakers and assign consistent labels to each.
• Segment speech with accurate timestamps.
• Identify the perceived gender of speakers.
• Flag audio-quality issues — unclear or distorted speech, strong accents, incorrect language, or synthetic/generated audio.
• Verify and correct pre-annotated segments when needed.

Requirements

• Strong listening skills and excellent attention to detail.
• Previous transcription or speech-annotation experience — this is explicitly wanted, not a "no experience needed" role.
• Native or near-native fluency in your target language/locale.
• A computer and stable internet connection.

Is there an exam?

Not described as a formal exam. Tasks are released weekly on a first-come, first-served basis per locale/batch — so consistent availability matters as much as passing a one-time test.

Pay

Not publicly disclosed. Same locale-gated pay structure as OneForma's other listings.

Who should apply

People who've already done transcription or speech-annotation work before and want to keep doing it, in a native/near-native language. Not the easiest first project if you've never transcribed audio professionally.

How to prepare

No exact TTA Academy course matches audio transcription specifically; AI Content & Digital Marketing is the closest general-skills overlap for content-quality review work (TTA Academy's own recommendation).

Part of the OneForma vendor guide. TTA Academy is not affiliated with OneForma.