search

ken107 / read-aloud

An awesome browser extension that reads aloud webpage content with one click
https://readaloud.app
MIT License
1.31k stars 227 forks source link

Add support for using ElevenLabs API please #404

Closed jkyndir closed 1 week ago

jkyndir commented 3 weeks ago

As the title says, it'd be great if voices from Eleven Labs can be used via its API.

ken107 commented 3 weeks ago

Could you do a rough estimate of how many articles per day you will be reading aloud? And from there calculate how many characters of text you will be synthesizing per day? Let's say a typical news article is about 1500 words or 6000 characters.

jkyndir commented 3 weeks ago

it depends, maybe 5-10 articles. but typically, i will skim stuff when reading aloud

ken107 commented 3 weeks ago

5 articles/day is about 30,000 characters/day, which is about 900,000 characters/month.

ElevenLabs Pro plan only gives you 500,000 characters/month. So you'll need the Scale plan which costs $330/month.

ken107 commented 3 weeks ago

The reason why ElevenLabs charge such a high price for their voices is that their target customers are businesses and enterprises, not end consumers like us. Business and enterprises would use these voices to synthesize marketing materials, voice overs, converting documents to audio, etc. They would save the synthesized audio files and play them back many times, thereby justifying the high cost they pay for synthesis.

We're not the target customers of ElevenLabs. For our use case, we synthesize the article and listen to it only once, then throw away the audio file. Hence the price appears absurdly high, for the very little value we receive.

jkyndir commented 3 weeks ago

You do have a good point. Elevenlabs is indeed pricey. I'm still trying to find an open-source/affordable TTS alternative that is easy to use and sounds really natural. Guess i can't have it all, lol.

BTW, dunno if it'd be possible to add support for this: edge-tts, which is free Microsoft Edge TTS. Their voices do sound nice.

jkyndir commented 3 weeks ago

@ken107 hi there, first of all, big thanks for all your hard work! Your extension is the only one out there that can read Google Play Books aloud from my library, which is simply amazing! You have no idea how much trouble you've saved me.

As mentioned in the comment above, I noticed Microsoft Edge offers free TTS voices that sound incredibly natural, almost as good as OpenAI/11labs. And it looks like these could be integrated into a Chrome extension. Here's an example I found: ms-edge-tts. But that extension is missing a lot of essential features and is really tough to use.

Any chance you could look into adding these Microsoft Edge voices to your extension? Thanks!

jkyndir commented 3 weeks ago

Just found out when installing your extension in the Edge browser, all those MS Edge natural voices become available by default, which is amzing!

But still, i wonder if it's possible to have them available in Chrome browser as well? since that's my main.

ken107 commented 3 weeks ago

391 I think probably possible, we'll check out and add it in a future release. Thank you for your feedbacks.