AI voice for ads: radio and video voiceover
A small business ad, a promo for a launch, a pre-roll on YouTube, a spot on local radio: most ads need a clear voice and a script that fits the seconds you have. Booking a voice actor for every version and every language takes time and money, and a flat robotic voice makes an ad sound cheap.
SuperClonica gives you natural voices in 31 languages, a free online test and a free engine that renders as many takes as you like. This guide shows how to fit a script to the time, voice it, mix it with music and which voices you may use in paid ads.
How many words fit in 15 or 30 seconds
These figures are measured with the engine at its default speed; read your final script aloud with the timer anyway.
- English at the default speed: about 3 words per second, so roughly 45 words for 15 seconds and 90 for 30.
- Spanish, German and French: about 80 words for 30 seconds; Portuguese about 75; Russian about 65.
- Leave a second or two for the brand name and the call to action at the end.
- If the script is too long, cut words rather than speeding up the voice.
Try your hook free online
Paste your opening line, 30 to 100 characters, into the page of any voice, choose the language and press Voice it. In a few seconds you hear whether the voice sells your offer, with no account. Try two or three voices before you decide.
Write a script for the ear
Write the web address the way it should be said, for example example dot com, because addresses are read in pieces.
- Put the offer in the first sentence and repeat the brand name at least twice.
- Write prices and phone numbers in words, for example nineteen ninety-nine: the engine skips currency signs and reads a phone number as one big number.
- Short sentences with one idea each; a period after each idea gives the listener a clear stop.
- End with one clear action: visit, call or download.
Voice, take and mix
For the final version, install the free engine with python -m pip install superclonica-cpu, put the script in ad.txt and run superclonica --text-file ad.txt --voice M2 --lang en --normalize --seed 2 --output ad.wav. --normalize sets an even level of −18 LUFS, and --seed gives another take of the same script: try several and keep the best. Then put the voice over music in any editor, lower the music under the voice and match the final loudness to what the station or platform asks for.
Which voices you may use in ads
For a paid ad, use the ten stock voices, M1 to M5 and F1 to F5, free under the OpenRAIL-M license, or your own voice trained in the SuperClonica studio from a recording of a speaker who agreed. Character and parody voices from the catalog are not for advertising, and never imitate a real announcer or a famous person. Check the ad platform’s rules on AI-generated content.