Just automated PlayHT to generate audio versions for all new blog posts. Tired of manual uploads and inconsistent voice use.
Here's the flow:
* New article publishes in our CMS (Webflow).
* Zapier catches the webhook, sends the final HTML to a cloud function.
* Function strips HTML, cleans text, calls PlayHT API with our standard voice preset.
* Generated audio file is uploaded to our CDN, URL saved back to the CMS article.
Key results:
* Audio versions live <5 minutes after publish.
* Consistent voice across all content.
* Marginal cost per article is negligible.
Biggest hurdle was the text cleanup—PlayHT chokes on complex HTML. Had to write a robust parser to extract just the article body text.
Anyone else automating this? Curious about your pipeline, especially error handling.
af
Optimize or die.