Can you tell what my app does by its App Store screenshots?
In my opinion, the goal of app store screenshots are to:
- Convey what the app does
- Define your value proposition (why would I download your app)
- Promote the brand of your app
Given the title, sub-title, and screenshots below, do you think I'm accomplishing my goal?
Title: Potes: Podcast Audio to Text
Sub-Title: Screenshot to Transcription
Screenshots: https://imgur.com/3X1oxJo
Potes is days away from beta (iOS only). If this sounds interesting to you, you can sign up to receive an email when beta launches at https://potes.webflow.io.
the biggest thing i’d change is the first frame. it should answer “what is this, and why should i care?” in one glance. right now the screenshots do some of that, but the value prop could be sharper and more specific to the actual use case. one non-obvious trick is to make the second frame answer a different question than the first, so you’re not repeating yourself across slides. i ran into the same screenshot-iteration problem and ended up building a tool for it, happy to share if useful.
Overall, I don't think you need that many images. One image that has more information is more than enough. For example, counting from the left, in the third image, you could have arrows pointing to the text saying "transcribe audio to text automatically" to tell the message.
Overall, I think it helps more overall to have a video that shows how your app works more than anything you can do with images.
The last three images are also good, since they say other things about the app.
As for your website, I am still confused about how your app works. For example, why is the screenshotting necessary? Can't I just download the entire transcript from the podcast after playing it?
I also don't see much of a use for generating audio at the time of the screenshot, wouldn't I want the transcript of the entire podcast?
I can see this useful for people that want to take notes on certain areas of the app, and that is a good niche use case, but if it can't transcribe the entire podcast, then I don't see it being too useful. But if transcribing the entire podcast is also possible, then you should state that on the website.
Overall, I think you have a solid simple design and I love that you're not a separate platform.
Speaking of websites and advice, I'd love to get a second pair of eyes once my product releases as well.
FYI, it's an AI API that allows users to host AI models on the cloud. Considering you're using whisper, we are thinking of adding a whisper API to our cloud in the future.
If I may ask, which Whisper API are you using? I searched around and I don't think OpenAI offers it as an API on their website.
If you're curious about my project, feel free to follow my twitter: https://twitter.com/TheRealEtch
Good luck on your product! Seems promising.
Hey RichardGao.
Thanks for your thoughtful reply.
I disagree on the amount of images. I think people mostly use the images on the App Store for helping decide whether or not they are going to download an app. I know I do.
Good point on the video. This is definitely a unique concept so a video might be wayyy more valuable.
The idea of screenshotting is to take notes. Instead of viewing an entire transcription of a podcast, it's, "I heard something I like. Let me take a note." The screenshotting kicks off just transcribing what you just heard. Maybe I need to make that clearer on my website and in my screenshots.
I'm thinking I change the text on the 3rd slide to "Potes transcribes the last 30 seconds from audio to text" to help signal that it's the screenshot that kicked that off.
Just followed you! Please do share when your twitter bio website releases. I get to a generic godaddy page now. Definitely interested in this space. I am using Replicate. It seems to be working pretty well so far. There are certainly costs but they are minimal.
Do you know if you'd be serving up something similar to Replicate? What sort of AI models are you looking at?
That makes much more sense, especially if it's a note taking app.
And yes, I am aware of Replicate. I am doing something similar, where you can run AIs off our API, and we're getting ready to release our stable diffusion API.
The difference is that we plan to be cheaper. Not by much, but the API costs add up. Replicate also does not have transparent pricing of how many AI generations it takes for their free trial to end (they say "you'll be asked to pay in a bit", but if you have paying customers already, I'm sure that doesn't matter to you that much.
Good luck on your product!
Very cool. I think we're still at the infancy of AI so all the power to ya. If you're cheaper and provide the same service I would consider a transition.
I don't have paying customers yet for Potes as I haven't launched it. Unfortunately I started to realize when I needed to use APIs like Replicate, that I'd incur costs. I was hoping to provide a more generous free tier, but have had to rethink that.
That's unfortunate. Maybe you can offer alternate pricing? For example, if a customer feels they won't use it enough for a subscription, you could have them use pay per use instead.
Doing this, you can offer them an "un-generous" free trial, where they'll just pay per use afterwards without the hurdle of signing up for a subscription they might not use.