Make longer content and convert subtitles directly using our API

Jul 28, 2023

We’re excited to announce several notable improvements to our text to speech API. With increased limits, new configuration options and support for directly converting subtitle files, Narakeet API now makes it much easier to create longer and more complex content.

Increased text to speech conversion limits

We have increased the limits for API jobs from 100 KB to 1 MB, so you can make ten times longer content directly using the long-content API. The new limits are available to all our commercial users immediately.

New configuration options

Until today, configuring the text to speech conversion jobs required inserting a script header. We’ve introduced two new query-string parameters to both the short content (streaming) and the long content (polling) APIs, making it easier to fine-tune the AI voiceover without any changes to the content. This makes it easier to pass existing files directly to our API for conversion.

• voice-speed selects the default voice speed. This has the same effect as setting the voice-speed script header
• voice-volume controls the default voice volume (loudness). This has the same effect as setting the voice-volume script header

Convert subtitle/closed captions directly

Lastly, you can now pass SubRip (.srt) and WebVTT (.vtt) subtitle and closed caption files to our API, and turn subtitles into realistic speech. A common use case for this function is to provide translated versions of audio tracks for a video. You can, for example, translate subtitle files from the original voiceover language into any of the 90 languages that Narakeet supports, then use the translated subtitle files in Narakeet to make an audio track.

This function was available only through the Narakeet web site until today, but you can now also integrate it into third-party applications, or your content automation workflows.

To convert a subtitle file automatically, just provide the MIME content type of the subtitle file in the Content-Type header when making the request, and upload the subtitle file in the request body. Both the short content and the long content API support this parameter.

For example, the CURL command line below will convert example.srt into audio, using the voice victoria and 10% faster than normal speed.

curl -D - --data-binary '@example.srt' -H 'Accept: application/octet-stream' -H 'Content-Type: application/x-subrip' -H "x-api-key: $APIKEY" "https://api.narakeet.com/text-to-speech/mp3?voice=victoria&voice-speed=1.1" --output output.mp3

For the code in Python, check out the example project on GitHub.

For more information, see the Converting Subtitles section of our text to speech API page.