Greg 🌱

Hey Product Hunters 👋 A lot has changed since our initial launch in 2021! Greg has helped folks care for over 1 million plants ( 🤯), we’ve expanded our team with some total rockstars, partnered with hundreds of *the coolest* plant nurseries around the world, and developed a community that has blossomed into one of the kindest and most helpful spaces on the internet. Now we’re stoked to introduce PlantVision! 🌱👀 With just a photo, PlantVision instantly unlocks *everything* (and we mean everything) you need to keep a plant happy and healthy. PlantVision combines the latest AR capabilities with our proprietary computer vision AI to instantly identify your plant’s: 🧬 Species 🍯 Pot height and width 🏠 Indoors/outdoors 🌤 Distance from a window (+ cardinal orientation coming soon!) In one tap, all that (+ more, like your locale) is synthesized to create a perfectly personalized care schedule for your individual plant. AND we’ll also tell you everything else you need to know, like: 💧 When (+ how much!) to water 🧪 If it should be fertilized (and when) 🐾 Is it safe for pets and kids? ☀️ How much sun it needs, and if it’s getting enough 💚 Is it better for beginners or veteran plantheads? 🪴 The best type of soil to use ➕ And lots and lots more! We also connect you to a whole community of folks growing your plant, which for some species tops 50,000 helping hands to field your questions and share tips & tricks. We know it can take a village sometimes! So, how did we do it? ⚙️ More than just a camera, PlantVision is built on a deeply optimized real-time scanning engine that we designed and built in-house to detect and analyze all the important information about any plant in any space – all within a fraction of a second! It combines the latest advancements in AR + AI to create a comprehensive model of the plant and its surrounding environment, which we input into our now-proven plant evapotranspiration model to generate an intelligent, context-specific care schedule. Our team built three separate custom PlantVision ML models that facilitate the detection capabilities. These models were intentionally built to be compact and efficient so in order to run them in parallel directly on the device. They include: 🔎 Plant Detector - locate and track all the plants in view 🏡 Scene Classifier - predict whether the plant is inside or outside 🧬 Species ID Classifier - Is this a golden pothos, or a heartleaf Philodendron? To make these models as accurate as possible, we need a LOT of data. Publicly available datasets (COCO, Open Images V6, etc) were amazing resources to get this off the ground, but could only get us so far. To push performance to the next level, we invested heavily into an in-house image labeling tool, where our resident plant experts could efficiently label, augment, and quality-control our dataset to make it bigger and more accurate. The investment certainly paid off! Our Species ID classifier maintains 95% precision across our top 500 indoor plants, and includes > 2,600 species, including popular hybrids and cultivars. As more people begin scanning more plants, we use that data to continuously update the models and increase these metrics. To bring it all together, the engine adaptively distributes CPU/GPU/ANE resources between the ML models and a series of 3D spatial analysis processes to provide ridiculously quick and accurate scans while still maintaining a cool 60 frames per second on the video feed – even on older devices! (For you iPhone Pro users out there, yes – we bring that LiDAR into the mix) Before you go 😸 Okay, that was a lot. Now is the perfect time to take a break (or a walk) and try it out for yourself! We wanna see screenshots of your coolest plant scans in the comments 👇 Thanks so much for being here and growing with us! Drop your username here, and we’d love to share a free lifetime of Super Greg as a thank ya. 🤗 And finally, an extra big thank you to @Chris Messina for the hunt! 🤩 Grow on, Alex and the Greg Team 🌱

Love love Greg — and indeed, he's helped me save my plants! This new AR experience for identifying plants is dope! Don't miss it.

This app first requires your email, then reports you have to pay for it. Shame! Deleted it without even free trial.