Launched this week
Local AI Art: PocketDiffusion
Private AI image generator — no cloud, no API , 100% offline
10 followers
Private AI image generator — no cloud, no API , 100% offline
10 followers
Most AI image generators rely on the cloud — your prompts, data, and creativity are sent to external servers. PocketDiffusion changes that. This is a **100% offline AI image generator** that runs directly on your iPhone using Stable Diffusion. No internet connection, no API costs, and no privacy concerns. Generate high-quality AI art anytime, anywhere — even in airplane mode. Built for creators who want full control, speed, and true privacy.
Offline generation is genuinely useful, especially for travel. One thing that would make this even better for me is a simple in-app history or gallery view so I can revisit past generations without losing them when I clear cache.
@asminy6i5 I built this first version to prove fully offline generation is viable on iPhone — now it’s about unlocking more advanced workflows like this.
Really appreciate you calling this out 🙌 , Thank You so much
Running Stable Diffusion fully offline on an iPhone is genuinely impressive — love that it's private and free of API costs. One suggestion: please add a way to import custom LoRA or checkpoint models from Files. Right now creators like me have tons of fine-tuned models on our devices but no way to use them, and expanding model support would make PocketDiffusion way more useful for serious workflows.
@sudenureliqx6g Great suggestion — LoRA / checkpoint import is definitely on the roadmap, just need to solve some Core ML constraints first.I will Implement in ASAP .
If you’re enjoying it, I’d really appreciate a 5⭐ review on the App Store 🙌
Running Stable Diffusion locally on an iPhone sounds impressive, love the privacy angle. One thing that would make this way more useful for me: a way to import my own LoRA or embedding files so I can fine tune the style to match what I'm actually working on, instead of being stuck with the base model output.
@rzakzldanf4h Really appreciate this — you’re absolutely right.
Right now PocketDiffusion is focused on making fully offline generation stable and accessible, but I know being limited to the base model isn’t enough for real workflows.
Supporting LoRA / embeddings from Files is something I want to add — the main challenge is adapting them to Core ML and keeping performance reasonable on-device.
Definitely on the roadmap as I push this toward more serious creator use cases 🙌
Offline generation is such a refreshing angle, and the airplane mode demo is a nice touch. One thing that would make this way more usable for me: a built-in history tab with thumbnails and the exact prompt and seed for each image. Swiping back to remember what settings produced a favorite result is the biggest friction point right now.
@necatipisk53626 i will keep more updates on it ASAP , Thank You so much for your feedback . It you like it , please give a review in app store .
finally tried this on a flight and was honestly shocked at how well it ran offline, the diffusion speed was smoother than i expected for a phone app
@nurcanhepmersin This made my day — “on a flight” is the ultimate test 😄
If you get a chance, I’d really appreciate a 5⭐ App Store review. It helps a lot as an indie dev 🙏
finally tried this on a flight and it actually generated solid images without internet, way faster than i expected for an offline tool.