Spokenly: Audio to Text AI app

I can’t remember the last time I left a review, but I feel compelled to do so. I have been looking for a Speech-to-text apps, and there are about a handful of them on the App Store. I tried each of them, and they were really good. Most of them have a payment model beyond a free trial, which is very fair. I gave Spokenly another try, and it is absolutely amazing. It does what the other apps do as well—convert speech to text—but being able to import your own APIs is a game changer. You can go into the app and, with a little effort, sign up for APIs from about 10 AI companies, and all of them come with a free quota. This makes speech-to-text almost free. I think the developer is doing us a huge favor, and I cannot thank him enough. I think having the ability to have “prompts”, which are customized dictations for different kinds of settings, is also a huge asset. Here, instead of any prefixed settings, you can truly define how you want that dictation to turn out, whether it is for legal, medical, casual, or other professional needs. For each one of them, you can set parameters for how you want them to look. Vadim, thank you so much. You have made a huge difference in my life.

This app is incredible. It lets you use a ton of local models, including the ultra fast and accurate Parakeet models from NVIDIA, Apple’s SpeechAnalyzer (which is actually different from the built in voice dictation, that is an older system), all for free and unlimited. And then if you pay, you can use more accurate cloud models from a bunch of providers and have an LLM rewrite and reformat it for different purposes. You can also directly provide API keys instead of the $10/month subscription. The UI is super polished, easy to use, does not require technical expertise. On iPhone, you can use the third party keyboard and have it auto start so when you press the globe button. On Mac, you can bind a key like right CMD or Option for push to talk, you release and it immediately pastes, or optionally simulates the keyboard inputs. Superwhisper and VoiceInk are also good, but they are both paid if you want a good experience, whereas this app gives you solid STT for free.

Right now, instead of typing this out, I am able to use my voice to dictate it. I use Spokenly all the time, every day, all day. The reason is because now, more and more, we're finding ourselves in LLM prompt windows, and you'd never want to start from scratch. You want to be able to fly at the speed of voice and not be limited by how quickly your fingers can move. So, Spokenly is amazing because it also has a local model. Right now, I'm using the Parakeet model by NVIDIA, which is a relatively small file, and it makes it so much better for those who are technical and want to take voice dictation to the next level. The reason is because a lot of times in the native voice dictation within Apple, it doesn't have the right technical words. aspects of the things you want to say.