Skip to main content

Formats & Languages – Precise Transcripts, Delivered Flexibly

A detailed overview of all supported input and export formats and the 50+ languages R2 Mechanics transcribes – including options for manual revision, interactive HTML outputs, and long-term archiving in WARC format.

Additional Details

The Web ARChive (WARC) format is an internationally recognized standard for the long-term preservation of digital content. It bundles transcripts, audio, and metadata into a single, self-contained archive file.
Why it matters:

  • Integrity: Ensures your data remains unchanged and verifiable for decades.

  • Compatibility: Widely used by libraries, archives, and research institutions.

  • Future-proofing: Guarantees accessibility even as file formats evolve.


AI Side‑Notes are AI‑enhanced contextual annotations embedded alongside your transcript. Depending on the language tier (Essential, Advanced, Expert‑Level), they may include:

  • Additional comments or clarifications for better understanding.

  • Cross‑references to external sources or related historical/biographical context.

  • Observations on tone, narrative coherence, and content plausibility for deeper insight.

In interactive HTML transcripts, Side‑Notes are accessible via hover or click, providing readers with a richer, layered context tailored to the level of analysis selected.


 Audio restoration enhances poor-quality recordings before transcription. This may involve:

  • Noise reduction to remove background hiss or static.

  • Volume leveling for consistent loudness across the file.

  • Equalization to improve clarity and intelligibility.
    This process ensures maximum transcription accuracy and provides you with a restored audio file that is easier to listen to and archive.


  Transcription Pipeline – From Input to Deliverables

Flexible Outputs for Every Use Case

R2 Mechanics delivers transcripts in formats tailored to research, archiving, and practical use. Whether you need a ready‑to‑read PDF, editable text, or a fully interactive HTML transcript with annotations and audio, our pipeline ensures maximum flexibility. For institutions and archives, we also provide WARC files – the internationally recognized standard for long‑term preservation – as well as restored audio for improved clarity and accessibility.

🎙️

Input Formats

  • 🎵Audio: WAV, MP3, FLAC, AAC, M4A, OGG
  • 🎥Video: MP4, MKV, MOV, AVI
  • 🎚️Multi-track recordings (interviews, panels, podcasts)
⚙️

Processing

  • ⏱️WhisperX / Whisper v3 JSON (timestamps & speaker separation)
  • 📝Side-Notes (annotations & contextual comments)
  • 💬Subtitle formats: VTT / SRT
  • 🎛️Audio restoration: noise reduction, EQ, volume leveling
📦

Output Formats

  • 📄PDF – Ready-to-read formatted transcripts
  • ✏️TXT / Markdown – Editable raw text
  • 🗂️JSON – With timestamps & speaker data
  • 🎞️VTT / SRT – For subtitling & media integration
  • 🌐Interactive HTML – Clickable timestamps, audio player & Side-Notes
  • 📦WARC – Long-term archive with transcript, audio & metadata

Supported Languages – Accuracy, Formats & Contextual Enhancements

R2 Mechanics transcribes in over 50 languages, offering different levels of accuracy, contextual enrichment, and archival-ready output formats.

Premium languagesEnglish, German, French, and Polish – receive expert manual revision and AI‑enhanced contextual annotations, combining very high transcription accuracy with deep historical, biographical, and plausibility insights.

All other supported languages are processed with the same secure, offline-first pipeline, but without manual expert revision. This leads to varying depths of contextual interpretation, as reflected in the High, Good, and Limited accuracy categories in the table below.

▼ Very High Accuracy (Premium)

▼ High Accuracy

▼ Good Accuracy

▼ Limited Accuracy (Input‑Dependent)