Audio and video transcription Transcription services: from audio to text.

What is a transcription?
Definition and common use cases

A transcription turns spoken content into text. In practice, that means an audio or video file becomes a formatted document with timestamps, speaker labels and, where needed, notes on non-verbal events. Read more in the Transcription glossary entry.

Common use cases: interviews for market research, court records, conference recordings, podcast episodes for accessibility and SEO, training videos and academic research interviews. In every case, a recording becomes a searchable, citable text.

Three packages, one outcome: your finished transcript
AI, Smart or Premium, depending on what you need

We offer three service levels. All packages accept common audio and video formats (MP3, WAV, FLAC, M4A, MP4, MOV, AVI, WebM and more). You receive the transcript as a Word file (.docx), and on request also as .srt (subtitles), .txt or .pdf. Full pricing information is also available on our prices page.

For everything at a glance, download the PDF: Transcription one-pager (PDF).

What is a transcription used for?
Typical projects from our customers

Why AI alone isn't enough
What human post-editing adds

Modern speech recognition can transcribe audio in minutes. On a clear recording with one speaker, accuracy runs at 90-95%. But technical terms, dialects, background noise, speaker changes and proper names remain difficult. In a medical interview, the AI might mishear "baseline" as "base light." In a legal context, "tort" can become "torte." That's why a well-maintained termbase pays off on transcription projects too.

That's why, in the Smart and Premium packages, native-speaker editors check every transcript during Medial Post-Editing (MPE). They correct technical terms, add timestamps, attribute speakers and apply the tolingo rule set (basic or extended transcription rules). The result: a citable, searchable text.

Basic and extended transcription rules
Which rule set fits your project

tolingo works to two standardised rule sets. You choose which one applies when you place your order.

Basic transcription

The text is captured exactly as spoken, including errors and filler words. Unclear passages are marked [inaudible], other languages as [foreign language]. Speakers are labelled "Speaker 1", "Speaker 2" and so on. No timestamps.

Extended transcription

Everything in basic transcription, plus: dialect colouring is corrected, non-verbal sounds (ums, stutters) are removed, listener acknowledgements are kept only where they add content. Notable events go in brackets (e.g. "(audio disruption)"). Sensible punctuation for long sentences. A timestamp before every speaker change. Pauses over 4 seconds marked with their length. Unclear words flagged with a question mark: (?landmark).

For a detailed practical guide with examples, see our guide: Avoiding mistakes in interview transcription.

Transcription and translation, one team
From the spoken word into any target language

In the Premium package, we deliver transcription and specialist translation as a single project. That saves you coordinating between a transcription provider and a separate translation agency: one project manager handles both.

A typical scenario: an international company has training videos transcribed in German, then translated into English, French and Spanish. The transcript becomes the source text directly for translation into 220+ languages. On request, we also produce subtitle files (.srt) for your video platform.