Your proper nouns, said the right way
One correction per line, in the form word equals pronunciation. WeDispatch applies it before every synthesis, across your whole catalogue, and the on-screen text follows the pronunciation so the highlighted word matches what you hear.
A synthetic voice always trips on the same words: a local official's name, a place, an acronym, a brand. One mangled word and the credibility of the whole audio takes a hit. Fixing it article by article, by hand, is not sustainable.
Your proper nouns, said the right way
Write your corrections
One per line, in the simple form word equals pronunciation. You keep your own lexicon, the one of words that matter to your newsroom.
Applied at every synthesis
The lexicon acts before every generation, including in AI charter mode: it is a deterministic rule, not an artificial intelligence guessing.
The text follows the voice
Highlighting and subtitles follow the corrected pronunciation, so the word shown on screen is exactly the one you hear.
Credibility preserved
Your place names, officials and brands are said correctly. A detail to the ear, but it is what reads as serious rather than amateur.
Consistent everywhere
A correction set once applies to your whole future catalogue. No more a word right here and wrong there.
No cost, no quota
The lexicon is a rule applied at generation, not an extra resynthesis. Fix as many entries as you need.
Charter-mode compatible
Because it is a rule and not a language model, the lexicon stays active even when you require zero AI on your texts.
Local and regional press (town, official and place names), sectors heavy in jargon and acronyms, and any brand that cares about being named correctly out loud.
Ready to give your articles a voice?
Book a 30-minute demo: we set WeDispatch up with you, no complex setup.
What is the format?
One correction per line, in the form word equals pronunciation. The whole word is targeted, the leading capital is respected, and the correction applies before every synthesis.
Does it use my quota?
No. It is a rule applied at generation, not a new synthesis. You can maintain a long lexicon at no extra cost.
Is it compatible with AI charter mode?
Yes. The lexicon is a fixed, verifiable rule, not a language model. It stays active even in charter mode, where no AI prepares your texts.
Does a new entry apply to older audio?
It applies to future generations. For an already-produced audio, regenerate it and the correction will be taken into account.