Documentation Index
Fetch the complete documentation index at: https://docs.beyondwords.io/llms.txt
Use this file to discover all available pages before exploring further.
Overview
BeyondWords offers voice cloning, so you can create audio that sounds like your journalists, podcast hosts, or other chosen speakers. All voice clones are powered by ElevenLabs. There are two types of voice cloning available:| Type | Description | Audio required | Generation time |
|---|---|---|---|
| Professional voice cloning | Generates high-quality, production-ready voice clones with the most natural and consistent results | 30+ minutes | Approx. 24 hours |
| Instant voice cloning | Quickly creates a voice clone from a short speech sample | 30 seconds to 3 minutes | Approx. 1 minute |
Professional voice cloning
Professional voice cloning creates high-quality, production-ready voice clones from longer recordings. It’s designed for the most natural, consistent results and best suited for publishing workflows where voice quality and accuracy are critical. Your BeyondWords plan includes a limited number of Professional voice clone slots. Each time you create a new clone, it uses one slot from this allowance.Create a Professional voice clone
Provide voice details
Click ”+ Custom voice” → “Professional voice clone” then enter voice details:
- Speaker: The name of the person whose voice you’re cloning
- Voice name: The name to assign to the voice clone
- Language and accent: This will be treated as the voice’s native language and accent and should match the language and accent used in the audio
- Gender: The voice’s gender (male, female, or neutral)
Upload audio
Upload 30 to 150 minutes of recorded speech. (See our recording guidelines and requirements below.)Once complete, click “Confirm”.
Record or upload the verification statement
The speaker will need to record or upload the verification statement to confirm they’re happy for their voice clone to be created and used.You can also upload files to support verification, such as voice cloning contracts or other written consents.Once complete, click “Submit.”
Set project access
Decide whether to make the voice clone accessible across all projects or specific projects in your organization.Once complete, click “Save changes”.
Wait for your voice to be cloned
It will take around one working day for the BeyondWords team to generate and review your Professional voice clone. We’ll contact you as soon as it’s ready.Your new voice will appear in your organization’s custom voices table and in voice selection menus across eligible projects.
Recording guidelines
- Record real content: Record the same type of content the voice will be used for. This helps train the model on relevant material and capture the desired tone, pacing, and delivery. We recommend recording each article individually to preserve natural structure and flow.
- Use a consistent delivery style: Speak in the tone, pace, and style you want in the cloned voice, and keep this consistent across all recordings.
- Protect voice quality: Take regular water breaks and rest your voice to maintain consistency. Rather than recording everything at once, record in multiple shorter sessions to reduce the risk of voice fatigue or strain.
- Use a quiet, consistent environment: Record in a quiet location and use the same equipment throughout. A professional studio is ideal, but you can also reduce background noise and echo with soft furnishings such as duvets, quilts, rugs, or curtains.
- Use suitable equipment: A good-quality external microphone is recommended. Built-in laptop or phone microphones may reduce the final quality of the voice clone.
- Reduce plosives: Use a pop filter to reduce harsh plosive sounds such as “p” and “b”.
- Keep microphone placement consistent: Maintain a consistent, comfortable distance from the microphone to avoid changes in volume or tone. Consider using a shock mount to reduce noise. Positioning the microphone slightly off-axis can help to reduce bursts of air and harshness.
- Minimize mouth noise: Mouth noise can be replicated in the voice clone and affect audio quality. To reduce this, stay well hydrated in the days leading up to recording and throughout the session. Limiting caffeine and alcohol can also help. If needed, chewing xylitol gum or eating green apple on the day may reduce mouth noise.
- Breathing and pausing: Natural pauses and breaths will be captured in your voice clone and can help it sound realistic. If you’d prefer a cleaner result, try to reduce breath noise by turning away from the microphone and keeping breaths low and consistent in volume.
- Make corrections naturally: If you make a mistake, re-record from a natural point. You can fix short sections without re-recording everything, as long as the result sounds consistent.
Recording requirements
Outlined below are the optimal recording requirements for your audio. If these can’t be achieved with your setup, you can use post-processing software to boost audio quality. Keep in mind that the output audio quality will sound very similar to the input audio quality.- File format: MP3 (192kbps+) or WAV
- Sampling rate: 44.1 kHz or higher
- Bit depth: 16-bit or higher
- Volume levels: Aim for -23 dB to -18 dB RMS, with peaks no higher than -3 dB
- Signal-to-noise ratio (SNR): Above 35 dB for minimal background noise
- Environment: Record in a quiet, echo-free space with consistent conditions
- Processing: Submit audio in its original, unprocessed form (no compression, limiting, or effects)
Instant voice cloning
Instant voice cloning lets you quickly create a voice clone from a short speech sample. It’s designed for fast setup and flexibility. Instant voice clones don’t use any cloning slots, so you can create as many as you like.Create an Instant voice clone
Provide voice details
Click ”+ Custom voice” → “Instant voice clone” then enter voice details:
- Speaker: The name of the person whose voice you’re cloning
- Voice name: The name to assign to the voice clone
- Language and accent: This will be treated as the voice’s native language and accent and should match the language and accent used in the uploaded audio
- Gender: The voice’s gender (male, female, or neutral)
Record or upload speech
Record or upload speech to base your voice clone on:
- Provide between 30 seconds and 3 minutes of audio (this can be across multiple recordings)
- Speak in the style you want the voice clone to replicate
- Keep your delivery consistent across samples
- Use clean, high-quality audio with minimal reverb, artifacts, or background noise
- Aim for audio levels between -23 dB and -18 dB RMS, with a true peak of -3 dB
Preview the voice clone
Use the preview tool to see how the voice clone sounds across different voice models, languages, and accents.
FAQs
What languages are supported?
What languages are supported?
You can create an Instant voice clone in up to 74 languages, depending on your chosen voice model: Afrikaans, Arabic, Armenian, Assamese, Azerbaijani, Belarusian, Bengali, Bosnian, Bulgarian, Catalan, Cebuano, Chichewa, Croatian, Czech, Danish, Dutch, English, Estonian, Filipino, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Hausa, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Irish, Italian, Japanese, Javanese, Kannada, Kazakh, Kirghiz, Korean, Latvian, Lingala, Lithuanian, Luxembourgish, Macedonian, Malay, Malayalam, Mandarin Chinese, Marathi, Nepali, Norwegian, Pashto, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Serbian, Sindhi, Slovak, Slovenian, Somali, Spanish, Swahili, Swedish, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu, Vietnamese, Welsh.You can create a Professional voice clone in up to 32 languages, depending on your chosen voice model: Arabic, Bulgarian, Chinese, Croatian, Czech, Danish, Dutch, English, Filipino, Finnish, French, German, Greek, Hungarian, Hindi, Indonesian, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Romanian, Russian, Slovak, Spanish, Swedish, Tamil, Turkish, Ukrainian, Vietnamese.
How many voice clones can I create?
How many voice clones can I create?
You can create as many Instant voice clones as you need, while Professional voice clones are subject to your plan allowance.If you need more slots, please contact support.