Nazar

iOS Developer

[Nazar @nazar_codes]

@decentricity @tmuxvim how is the experience so far?

[Nazar @nazar_codes]

@RhomasDaThomas @jacobrodri_ it’s a web funnel detached from the app store app binary. apple has no say in that

[Nazar @nazar_codes]

This and 100k MRR

[Dan Hollick @DanHollick]

Summer in the garden office.

[Nazar @nazar_codes]

@Swaroop_1102 i’m so happy for you 🥹

[Nazar @nazar_codes]

@maverickecom live

[Nazar @nazar_codes]

No more endless scrolling to find the right icon

[Daniel Storm @DanielStormApps]

This is such a huge QOL improvement!

[Nazar @nazar_codes]

Kids these days will never experience the thrill of doing random shit until the error disappears
#wwdc26

[Nazar @nazar_codes]

@simanerush @_julianschiavo It’s always a pleasure to learn from your videos

[Nazar @nazar_codes]

@fayhecode hey, are you sure “wellbeing sanctuary” is a set of keywords that people search on appstore?

[Nazar @nazar_codes]

@affprinter slideshow

[Nazar @nazar_codes]

@lucaspatiri_ UGC

[Nazar @nazar_codes]

@ugcmaxxr @adriansolarzz slideshow

[Nazar @nazar_codes]

Today I polished the prompt I give to the AI and added some guardrails so users won’t get their homework done with the help of my app 🙂
Also thinking about giving the AI context about the user, so it could give more personalised responses

[Nazar @nazar_codes]

Finished the implementation of hybrid approach. Settled on gemini-3.1-flash-lite, seems to be capable enough and dirt cheap.

Funny thing is, I’m actually throttling the stream on the UI to make it look like it’s running at 110 t/s, because the actual response is near instant

[Nazar @nazar_codes]

Finished the implementation of hybrid approach. Settled on gemini-3.1-flash-lite, seems to be capable enough and dirt cheap.

Funny thing is, I’m actually throttling the stream on the UI to make it look like it’s running at 110 t/s, because the actual response is near instant

[Nazar @nazar_codes]

Don't know what model to choose—needs to be relatively capable and dirt cheap. Any suggestions?

[Nazar @nazar_codes]

Don't know what model to choose—needs to be relatively capable and dirt cheap. Any suggestions?

[Nazar @nazar_codes]

Been implementing a generic AI chatbot for my iOS app. First I wanted to go with a privacy-focused approach with local LLMs, but Apple Intelligence is only available on iPhone 15 or up. Then I tried SpeziLLM - a library that lets you run models on device via XML…

[Nazar @nazar_codes]

The trade-off is that users have to download a 2-4GB model, basically killing user retention. So I decided to go with a hybrid approach: check if an Apple Intelligence model is available, if not I switch users to an API-based one, while saving no metadata or analytics whatsoever

[Nazar @nazar_codes]

@FloWritesCode thanks, I’m literally finishing building chat bot view for my app

[Nazar @nazar_codes]

@azamsharp what tech stack would you recommend for a mobile developer?