... been building AyniCam independently from Morocco. The idea was initially simple: Can an iPhone become an intelligent dashcam using its camera and on-device AI? That question gradually turned into: camera detection segmentation tracking incident recording navigation 3D maps There have been plenty of failed models, overheating phones, strange Core ML behavior and experiments that looked much smarter at 3 AM than they did the following morning. But the app is now live and continuing to evolve. One thing ... ... love to learn from the Product Hunt community: What would make you actually use an AI dashcam

... Naoufel, a solo dev from Oujda, Morocco. I built AYNI Cam an iPhone app that turns the phone into an AI dashcam: real-time road detection, incident recording, all on-device, no cloud, no account. I'm porting it to Android now, and the platform gap is bigger than I expected. On iPhone, Core ML runs my YOLOX model on the Apple Neural Engine and I get ~20 FPS at 640 640. One API, works on every device. On Android ...

... Since launching on ProductHunt we've been doing a lot of thinking. Conversations with fleet managers, feedback from this community, and honest reflection on what detectly actually is versus how we were describing it. We got the framing wrong. detectly isn't another fleet safety tool. It's the layer that sits between dashcam

... automatically detects near-miss incidents from dashcam footage such as tailgating, cut-ins, hard braking, pedestrian hazards, lane drift and gives every driver a risk score. No new hardware needed, works with any existing dashcam

... What a day. detectly officially launched on ProductHunt today. Quick recap for anyone just finding this. detectly automatically detects near-miss incidents from dashcam footage. Tailgating, cut-ins, hard braking, pedestrian hazards, lane drift. Every incident gets an annotated clip and every driver gets a risk score. No new hardware needed. I built the whole thing solo over the past few weeks, the ML pipeline, the dashboard, billing, everything. Today s been about getting it in front of as many ... ... build it! The human side of that workflow is probably harder than the ML. Comment from Kunal Sharda(@kunalsharda): Congrats, Kieran. Solo across the ML pipeline, dashboard and billing in a few weeks is genuinely impressive. And the reactive-vs

... believe or not. Lack of accountability: Without transparent records, we can't even conduct post-mortems when AI makes mistakes or causes harm. It's like being in a race car with a welded accelerator and no brakes or dashcam. Exciting, but deadly. So, the three of us a developer, a philosopher, and a product builder decided to do something about it. We wanted to equip AI with "brakes" and a "black box." This was the beginning of Oracle Ethics ...