Create an audio guide BETA
Enter the text for each stop, choose a language and voice – we turn it into audio files you can download and use in your museum.
Create an audio guide BETA
Enter the text for each stop, choose a language and voice – we turn it into audio files you can download and use in your museum.
How the audio guide generator works
With the audio guide generator, museums, memorial sites and exhibition venues can create audio files for their tour free of charge – no recording studio, no voice actors and no sign-up. Enter a text for each stop, choose a language and voice, and download the finished files.
The files work for audio guide devices, QR codes next to exhibits, your own website or an app. Each stop is saved as a separate file with its number and title, or all of them together as a ZIP.
Features at a glance
- Free and no account needed
The generator is free to use and requires no sign-up. It is still in BETA – we keep improving it.
- Everything stays in your browser
Speech is generated right on your device. Your texts are not sent to us; stops and settings stay saved in your browser until you delete them. Only the voice itself is downloaded once the first time you use it.
- 56 voices in 18 languages
From German, English and French to Czech, Swedish or Greek – with a preview before generating and a quality rating.
- Adjust the delivery
Set the speed, the pauses between sentences and paragraphs, and the speaking rhythm (even or lively).
- Names pronounced correctly
Foreign names such as “Château de Pierrefonds” are detected and spoken in their own language. You can mark individual passages, and the pronunciation list lets you define how a word should sound – across all stops.
- Numbers and years
Years, ordinal numbers and common abbreviations are spelled out, so “19th c.” becomes “nineteenth century”.
- MP3 or WAV
Up to 20,000 characters per stop and as many stops as you like. MP3 for small files, uncompressed WAV for further editing.
Tips for good results
- Short sentences and a blank line between paragraphs make the text easy to listen to.
- Listen to every file once before using it – especially names and foreign words.
- The voices are computer-generated; it is best to let your visitors know.
Licences and sources
Speech is generated with free software: Piper (MIT) · eSpeak NG (GPL-3.0) · ONNX Runtime (MIT) · lamejs (LGPL-3.0)
The 56 voices come from the collection rhasspy/piper-voices. Most voices are in the public domain (CC0, Public Domain) or licensed under Apache-2.0 or the Unlicense.
These voices are licensed under CC BY – please credit the voice name and licence when you use them:
- cs_CZ-kasandra-medium (CC BY 4.0)
- de_DE-mls-medium (CC BY 4.0)
- en_GB-alba-medium (CC BY 4.0)
- en_GB-aru-medium (CC BY 4.0)
- en_GB-vctk-medium (CC BY 4.0)
- en_US-libritts_r-medium (CC BY 4.0)
- es_ES-sharvard-medium (CC BY 3.0)
- fr_FR-siwis-medium (CC BY 4.0)
- fr_FR-mls-medium (CC BY 4.0)
- it_IT-serena-medium (CC BY 4.0)
- it_IT-serena-high (CC BY 4.0)
- nl_NL-mls-medium (CC BY 4.0)
- sl_SI-artur-medium (CC BY 4.0)