# Trying to use a text to speach api

**URL:** <https://community.thunkable.com/t/trying-to-use-a-text-to-speach-api/3241608>\
**Category:** Web API's\
**Tags:** webapi\
**Created:** [October 1, 2024, 10:18pm UTC](https://community.thunkable.com/t/trying-to-use-a-text-to-speach-api/3241608 "2024-10-01T22:18:23Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![thefantabobworlddne](https://avatars.discourse-cdn.com/v4/letter/t/ea5d25/32.png) [@thefantabobworlddne](https://community.thunkable.com/u/thefantabobworlddne)\
**Post date:** [October 1, 2024, 10:18pm UTC](https://community.thunkable.com/t/trying-to-use-a-text-to-speach-api/3241608/1 "2024-10-01T22:18:24Z")

</div>

Hi,  
Im having a problem finding a way to use a text to speach api. the deffault text to speach blocks do a lots of error in some languages like in french (when you have a ‘-’ like in “comment vas-tu”, it skip the word after the ‘-’ and just say “comment vas” and some prononciation mistakes). ive search in others topics but the way to solve this problem doesnt work with me or are not realy usable in my case. i need to be hable to play the sound with very low latency cause its chat language app. does anyone know a way to do that ?  
by the way im triyng to use openai’s tts-1 model and thunkables is not able to play the receved sound.  
Thanks in advence and sory for my bad english (tell me if something is not clear or understandable)

---

<div class="post-metadata">

**Author:** ![joemoukarzel88skm4pi](https://sea1.discourse-cdn.com/flex015/user_avatar/community.thunkable.com/joemoukarzel88skm4pi/32/185210_2.png) [@joemoukarzel88skm4pi](https://community.thunkable.com/u/joemoukarzel88skm4pi)\
**Post date:** [October 2, 2024, 6:22am UTC](https://community.thunkable.com/t/trying-to-use-a-text-to-speach-api/3241608/2 "2024-10-02T06:22:54Z")

</div>

Unless you find an TTS API service that outputs a URL, you’re going to be in a world of pain. I have tried Deepgram, Google TTS, Azure TTS, IBM Watson and ResponsiveVoice. They all seem to output a binary of wav or mp3. I wasted countless hours trying to figure it out.  
Apparently there is a solution using backend integration and switching between StP and DnD but i couldn’t get it to work:

> [@Play mp3 sound from API response](https://community.thunkable.com/t/play-mp3-sound-from-api-response/2183817):
>
> I am using an API POST call and it’s returning an mp3 file. This works fine in Postman and plays the sound but not in Thunkable Live. Can I just attach the green response block to the play block and have it play the sound? I’ve tried that and it’s not working and I’m not sure what else to try. I included the get object from JSON block and that didn’t work either. This topic makes me think it’s probably not possible: [Download mp3 and play via Sound](https://community.thunkable.com/t/download-mp3-and-play-via-sound/1440164)

> [@ElevenLabs TTS Posting, Web API](https://community.thunkable.com/t/elevenlabs-tts-posting-web-api/3011510/4):
>
> this is how I am setting the blocks, would this be correct? I am getting an error from my live editor.

---

<div class="post-metadata">

**Author:** ![joemoukarzel88skm4pi](https://sea1.discourse-cdn.com/flex015/user_avatar/community.thunkable.com/joemoukarzel88skm4pi/32/185210_2.png) [@joemoukarzel88skm4pi](https://community.thunkable.com/u/joemoukarzel88skm4pi)\
**Post date:** [October 2, 2024, 6:58am UTC](https://community.thunkable.com/t/trying-to-use-a-text-to-speach-api/3241608/3 "2024-10-02T06:58:07Z")

</div>

OK! I found a TTS API that can output an MP3 as a url. It’s called Unreal Speech:

 ![Unreal](https://us1.discourse-cdn.com/flex015/uploads/thunkable/original/3X/f/9/f93344726c69ec47c9140e497a0639f71b5c75b5.png)

Unfortunately it doesn’t support French at the moment, so it won’t be much help for your use case, but I thought I would post it as alot of people have been asking for a simple method for high quality TTS that’s easily integratable in Thunkable.