Simple API included - take my money (updated)
This review is updated due to a change (correction?) in the API...
I've purchased a few text-to-speech services both from AppSumo and externally. Lazybird is more or less a wrapper over all of the standard 'non-ultra realistic' models out there. More than AWS Whisper, but not as many as some other services. As the developer says, there is no voice cloning at this time - that is a damn computationally intensive service, and I would not expect it to be included in a LTD. The result is that one the surface, Lazybird offers not much more than the dozens of similar products in terms of voices. As such, it is miles away from the ultra-voice quality of PlayHT and Elevenlabs. But it's also a LOT less costly, even without the LTD.
The project and design board is simple, to the point and effective. For folks wanting to create video pops or B-tier ebooks, it will be a breeze.
However, the reason I grabbed this offer as they have a simple API included with the plan.
My mega-investment in a LTD with PlayHT turned out NOT to support APIs in the end, so that entire service was shelved. Other text-to-voice wrappers also exclude API.
Lazybird offers API on both tiers being offered, with the same cost per character limits as working with the service via the project dashboard. Full marks there.
The speech API can handle plain text as well as SSML (Speech Synthesis Markup Language) which means that you can write apps that define not only what is said, but HOW it is said.
Generating the API key for your account is very simple.
Within 10 minutes of buying Lazybird, I had implemented it into to my make.com workflow, switching our Elevenlabs monthly costs with a LTD. While the voices are not quite as human-sounding, for my needs (e.g. my voice based snarky secretary) it is good enough.
This product is a good example of KISS.
** UPDATE **
Unfortunately, over the few weeks of using Lazybird, I ran into a few issues. First, the SSML feature was modified (I suspect to avoid passing model-specific SSML requests to models that would crash if they did not recognise them.) Rather than writing code to 'parse out' model-specific SSLM to modules that would not support it, Lazybird now severely limits the use of emphasis and other important in-speech features. I was able to live with that.
Then suddenly, all my API calls all started crashing. It took a while, but eventually, I discovered that the API calls were being rejected due to text/SSML exceeding limits. Given that my generation of content had not changed, I contacted support (who are excellent, BTW). They informed me that the *character* limit for SSML and text is now 2000. Alas, that is a deal breaker for me. Even a short daily agenda summary is not possible with such a limit, let alone reading emails, etc. Sure, you can pass the speech in batches, but stitching the various audio files together is a process I am not willing to code. As a result, I requested a refund.
Even so, I still give LazyBird 5 stars. The API inclusion still makes this a terrific deal. But with the changes to limits after the initial launch, it is not longer suitable for my needs.
So, if you have only basic, small needs and no need to voice variations/emphasis in your output then Lazybird is still excellent value. If you are a power user, looks like we are back to using ElevenLabs.

Ellis_Lazybird
Jan 10, 2025Such an awesome and insightful review! Thank you so much for this!
You hit the nail on its head there, keeping it simple is exactly our philosophy working on this product. We'll add some more powerful features and higher quality voices but as you said, they are inherently expensive and can't be included in the LTD deal if we want to be sustainable.