imgproxy

Maker

The iOS app is built with cutting-edge mobile technologies: Swift as our main language, Swift UI and UIKit for user interfaces, Lottie and Motion for complex animations, BBMetalImage and MetalPetal for image and video processing along with native Apple SDKs, Alamofire and GraphQL with Apollo for API, Fastlane and CircleCI for release management and continuous integration, Lokalise for localization management, Firebase and Amplitude for analytics, AudioKit and rnnoise for audio quality, Core ML to embed our machine learning models into the app. We applied some Machine Learning algorithms (built with Pytorch, OpenCV, ONNX, Core ML, and other libraries and tools) to work smoother, smarter, and safer for users and embed more sophisticated features, such as NSFW and restricted content recognition, spoken language identification, speech-to-text, multilingual text classification based on users’ interests, and those to adapt sound and video quality, ranking, and interests classifications. The app's backend is based on Ruby on Rails with GraphQL for communication with clients via API, Elasticsearch to search for content and users, and Timescale DB to store stats and users' ratings. Everything is deployed to a Kubernetes cluster with tasks to assign and report via Faktory. We also used some widely known open source products from Evil Martians: AnyCable Pro to arrange real-time communication and imgproxy to scale avatars and cover photos.

Report

Share

4yr ago

Fuat Bikbulatov

FEED

Maker

Hey Product Hunt! We are FEED! Our project and our team is not quite your average or normal one. We are a mix of directors, designers, creatives, & producers from the advertising industry and the IOS developers gang - altogether a similitude of Evil Martians. Over the past 10 years that we’ve been working in digital marketing, we noticed that when it comes to content - everything is focused on individuality, it’s all about one person’s personality and very often too self-centered in its approach. So we decided to give it a shot at creating a product that helps take content creation back from the individual focus to the collective collaborative space. This is pretty much what FEED is all about. Users can communicate to each other with short video-messages and, as a result, create a collective film together. This film is edited by special AI tools, and can be shared in the user’s profile (and across other social platforms too!) This Film - is a collective virtual “property” - it’s a piece that connects people from all across the globe. It’s the ultimate community builder solution. The process of communication in the FEED app (all video, no writing) helps people to visually connect with others online and to truly feel and understand the point of view of the other people in that particular live conversation. The video-format makes the conversation personalized, in contrast to texting. We worked hard to integrate a lot of different features to give these chats a twist, to make it easy to communicate and most importantly - to make it fun! Video walkie-talkie recording mode, for example, enables the user to send a video message with just one tap of a button in the app. For those who want to step-up their video content game, we added the video-editor, which has tons of different effects inside to play around with. We also added a built-in music assistant who automatically adds a mood. In the feed there is no more swiping up - you can now swipe in any direction! There are personal daily user logs, so you can keep up-to-date on participation. And with our sophisticated algorithms we can share your video with other users. Today marks 1.5 years since we started building FEED, and now, it’s time to share our project with the world. We invite you to press the JOIN button and try it out! We look forward to hearing from all of you hunters, please do share all your great ideas and comments in here, and we’ll be happy to answer all of your questions!

Report

Share

4yr ago