Omni Vox

Voiceover

A voice that holds for six hours, not six sentences.

Narration and dubbing in seven launch voices, including fuṣḥā and Gulf Arabic, delivered as podcast-ready masters levelled to spec — not a folder of clips for you to assemble.

The hard part is the long form

Any tool can read a sentence well. The difficulty is an audiobook: keeping the same voice across six hours, not drifting in pace between chapter two and chapter nine, pronouncing a name the same way every time it appears, and holding a consistent loudness so a listener never reaches for the volume. That is a production problem more than a synthesis one, and it is where most of the work goes.

How it runs

Script to master

The pipeline that turns a book into something publishable.

  1. 01

    Prepare the script

    The text is split into speakable chunks at sentence boundaries, with pauses, emphasis, and pronunciation marked. Names and terms get a pronunciation table so they never drift.

  2. 02

    Handle the passages that must not be synthesised

    Qur'anic recitation defaults to a reciter's library rather than a synthetic voice: verses are detected, matched, and spliced in with the loudness matched. You can also supply your own recording.

  3. 03

    Generate and stitch

    Chunks are generated with continuity conditioning so the voice does not reset between them, then stitched with natural room between sentences.

  4. 04

    Master to spec

    Each chapter is levelled to podcast loudness (−16 LUFS stereo), de-essed, and checked for clipping. This is the step that makes it sound produced rather than generated.

  5. 05

    Check it against the script

    The finished audio is transcribed back and compared to the script, so a dropped clause or a mangled name is caught by machine before it reaches you.

What you receive

  • Podcast-ready MP3, levelled to −16 LUFS
  • One file per chapter, plus a single joined master
  • WAV masters on request
  • Dubbed video with the original timing preserved
  • Burnt-in subtitles, rendered right-to-left correctly
  • The pronunciation table used, so future volumes match

The voice catalogue

Find the voice for it.

Filter by accent, gender, age and style, then pick a language and hear it read one of our lines. Every voice can read every language.

Read in:
Filter21 voices
Reading:

“Every language deserves to be heard as clearly as the first one.”

  • Bella

    Female · Middle Aged · American

    Professional, Bright, Warm

    ProfessionalInformative Educational

  • Roger

    Male · Middle Aged · American

    Laid-Back, Casual, Resonant

    ClassyConversational

  • Sarah

    Female · Young · American

    Mature, Reassuring, Confident

    ProfessionalEntertainment Tv

  • Laura

    Female · Young · American

    Enthusiast, Quirky Attitude

    SassySocial Media

  • Charlie

    Male · Young · Australian

    Deep, Confident, Energetic

    HypedConversational

  • George

    Male · Middle Aged · British

    Warm, Captivating Storyteller

    MatureNarrative Story

  • Callum

    Male · Middle Aged · American

    Husky Trickster

    Characters Animation

  • River

    Neutral · Middle Aged · American

    Relaxed, Neutral, Informative

    CalmConversational

  • Harry

    Male · Young · American

    Fierce Warrior

    RoughCharacters Animation

  • Liam

    Male · Young · American

    Energetic, Social Media Creator

    ConfidentSocial Media

  • Alice

    Female · Middle Aged · British

    Clear, Engaging Educator

    ProfessionalInformative Educational

  • Matilda

    Female · Middle Aged · American

    Knowledgable, Professional

    UpbeatInformative Educational

  • Will

    Male · Young · American

    Relaxed Optimist

    ChillConversational

  • Jessica

    Female · Young · American

    Playful, Bright, Warm

    CuteConversational

  • Eric

    Male · Middle Aged · American

    Smooth, Trustworthy

    ClassyConversational

  • Chris

    Male · Middle Aged · American

    Charming, Down-to-Earth

    CasualConversational

  • Brian

    Male · Middle Aged · American

    Deep, Resonant and Comforting

    ClassySocial Media

  • Daniel

    Male · Middle Aged · British

    Steady Broadcaster

    FormalInformative Educational

  • Lily

    Female · Middle Aged · British

    Velvety Actress

    ConfidentInformative Educational

  • Adam

    Male · Middle Aged · American

    Dominant, Firm

    Social Media

  • Bill

    Male · Old · American

    Wise, Mature, Balanced

    CrispAdvertisement

Questions we actually get

Can you use my voice?

Yes, with a recorded consent step and a signed licence stored against the voice. We clone end-users' voices, not stock personalities, and the consent is a requirement of the providers as well as of us.

How do you handle Qur'anic recitation?

By default we do not synthesise it. Verses are detected and spliced from a chosen reciter's library with the loudness matched. You can supply your own recording instead. AI voicing of Qur'anic text is available but is never the default.

What if one sentence comes out wrong?

Individual sentences can be regenerated and dropped back into the master without re-rendering the book. Fixing a line is a minute's work, not a re-run.

Which Arabic do you narrate in?

Fuṣḥā for most published work, and a Gulf-accented voice where the audience expects it. The catalogue is on the calculator; you can hear each one on this page.

Do you dub video as well as narrate books?

Yes. The audio is translated, generated, time-fitted to the original, and muxed back. Subtitles can be burnt in, with right-to-left rendering handled properly.

Price your recording.

Upload the script and the calculator estimates the finished hours and gives you the number.

Get an instant quoteSee the rates