I really wanted to have background audio and make it so the keyboard would directly record audio etc, but my first pass didn't make it through app review (and that was just keeping background audio listening AFTER you'd already started a recording). I could maybe have fought it, but figured if I was already butting up against app review there was little point as they'd likely reject in a future release anyway.
Re: analytics, It is quite weird having no idea how many people are using the app. It does leave you a little blind, but I figure people will get in touch if they have big enough problems with it.
For the Google app, I believe Gemma E2B and E4B both have audio input, so I suspect they're using one of those.
I built my own because I was frustrated with a lot of the free options. Largely because a lot of them had an upsell to be able to do the secondary post-processing step with an LLM. And it wouldn't pick up things like emojis properly or say numbers. Because of that, I left quite a lot of options in there for customising and adding additional steps, etc. Feel free to take a look: dictator.robgough.net
My initial Mac version actually had three additional steps that you could toggle, obviously at the cost of some speed. That is what the website talks about, although nowadays for my own use I've reduced that to just one step and found that it's pretty great. I've got a new version in test to tidy that up, but still lets you define as many steps as you want.
I think that's parakeet doing its job there. That is a closer reflection of what you've actually said. The trick is then throwing that output through some additional deterministic and non-deterministic steps to tidy it up however you prefer. It's exactly what I do with my free and open source dictation app (dictator.robgough.net) for Mac+iOS. And of course, everything stays entirely on-device. Gemma E4B is really wonderful for that second step, it's great at language – but takes up 6-7GB RAM.
It pairs quite nicely with agentic development as it has a history of plenty of open-source projects published on GitHub, which means they have learned to work with Rails et al. rather well.
It also helps to have a "boring" framework with strong opinions and strong community standards etc.
I wouldn't claim it to be the best as I'm not sure how you'd measure that, but I can say that in my experience it is letting me build things to a decent standard really rather quickly.
If you're building something new today, I'd generally recommend starting with a framework that you already know. But for those of us who already know Rails, it continues to be a wonderful choice. I'm playing with Phoenix LiveView for some projects, which is letting me build real-time UI's really easily – but they have some real-time requirements which Rails can do but is not it's strength. For anything a little more CRUD, it's a no-brainer.
What makes you think your usage is getting cut by 50% post July 13th? It's just that you can use up to 50% of your usage on Fable right now, and that is what is going away.
I'm going to miss Fable too, I found it surprisingly tough going back to Opus when we lost Fable first time round, but paying API costs is simply out of the question for me right now.
Quickly switching between multiple sessions, and seeing which sessions are waiting for input.
I generally run a session per project. Then within each session, a window for Claude, a window for running whatever dev server/logs, a window for neovim, and finally a plain terminal window for things like kicking off deploys etc.
I'm a big tmux fan, but admittedly I've found myself switching between a lot of sessions recently and wishing it was a little bit more mouse friendly.
If you're a Mac user, and can forgive the shameless plug, you might find belfry.robgough.net useful. Connects to local and remote machines entirely through tmux, ssh and libghostty - populating the sidebar from tmux session info, and optionally adds a little Claude visibility in there for good measure.
There's even an iOS/iPadOS version – though for the moment you'll need to build that yourself. Source is all on Github.
I've been surprised by how far my motivation has taken me. I have multiple projects now that are so much further ahead than I have ever managed before. I would get stuck on some small issue or key-decision and then would struggle to move past it.
Certainly, I still have those tendencies, but it's easier than ever to push through and build a throwaway version with code that might not be what I end up shipping, but is enough to help me understand the final shape of the solution. Once that happens, I'm unblocked and away I go again.
The mental cost of building throwaway code has essentially dropped to zero for me. Of course, I keep working now until light mode comes on automatically on my Mac, but that's a whole other problem.
They've said there are limits, and increased limits for those on iCloud+ ... so it seems that Apple is in the selling LLM access game now. I don't think there are any details yet on the nature of those limits, and whether they can be increased as required etc.
I'm fully expecting an updated foundation model of some kind, but I would bet money they don't utter the words "Google", "Gemini", or "Gemma" even once.
Mac only I'm afraid, but I already did [1]. Packaged it too as I figured it might be useful to others [2] (and I'd want to install it on machine's that I might not have Xcode on)
I spent the best part of a couple weeks making improvements and tidying up the UI, but to actually get something working was essentially only a couple of prompts.
I've got a home-built dictation app that uses a local model to clear up the text and fix grammar. It was super easy to build. I’m extending it to capture meeting notes and summarise too. All on-device.
I saw a little app the other day, I think someone posted on here, that looks at your screenshot and renames the file based off the contents of the file.
There's tons of little examples like that. For a lot of use cases, you really don't need the frontier models.
I recently built a similar app, and so hit the same limitations – I wasn't too upset on Mac, happy to distribute without the App Store (though it's a shame).
Where I was more frustrated was how much this limited the potential usability of the iPhone app. Because of app store restrictions it is a far worse app ... though like in your example, still useful to a degree.
I can only hope they use the new CEO as an opportunity to seriously re-evaluate their entire approach to how they work with developers, though I'm not actually expecting them to. If anything, with the increase in apps being created via AI tools I worry they will go the other way.
It's not as powerful as Gemma 4, but I think they likened it to GPT-3. It's perfectly capable of looking at images and classifying them at the level you'll need for this app. And it runs everything on the Apple Neural engines, so decently quick. Of course, this assumes that your users are using Apple Silicon processors, I believe that's the limitation – and they must have enabled Apple Intelligence which downloads the model at that point.
As clever as this is, it seems like the names are fairly straightforward (as you'd want!) – did you try using the on-device Apple Foundation model at all? That's actually pretty powerful for a use case like this, and if you're happy to require the user has Apple Intelligence turned on already, your shipped app can end up being tiny. The biggest concern for an app like this is how much RAM you end up using trying to run it. Especially if we end up with lots of different apps all doing the same thing.
Being able to super-power apps with on-device models is a lot of fun. I recently did the same building my own dictation app using small local models, and I still can't believe how effective it is. The download is just 20mb, though it will download parakeet ~475mb for audio, but can use the on-device model as the second-pass LLM and works pretty well (though better models are available to download and use e.g. Llama 3.2 4bit and Qwen 2.5 7B 4bit)
I'm currently building a little tool for a professional photographer friend to go through and classify images in their photoshoots, so I can build a searchable db for them to quickly find very specific images in the future. I simply don't think it would have been possible for me to build a tool like that just a couple years ago at any price.
Review times have already spiked, apparently, although I've never found them particularly good. There used to be a third-party website that tracked this via user submissions, though that was shut down when they made genuine efforts to improve timings several years ago as it was deemed no longer neccesary.
I think it's a fair criticism that there will be a lot of people putting "lower-quality" apps on there. I do think Claude did a better job of this than I would have managed. It certainly seems to work well enough.
I do believe Apple will have to rethink exactly what they want to gatekeep and why. If nothing else, in the context of allowing people to get more out of their powerful devices. For example, I wish my iPad Pro could actually do more useful things with it's M4 processor. Hopefully that's something the new CEO has on his list.
I think that's a fair question. For this app I didn't see the harm in releasing it. You have to pay for the account etc. to get it onto TestFlight anyway, so I might as well just put it out there. And avoids me having to resubmit every 90 days. There also wasn't a lot of custom "me" stuff in this one, but I could see going that route for other apps in the future.
[ my public key: https://keybase.io/robgough; my proof: https://keybase.io/robgough/sigs/B9QUm5q3cFdtCQFnvSl82vQxy8ua7GFsnejCuicA9XM ]