About Roboflow
Roboflow is a computer vision platform that turns images, video and live streams into structured data. Teams use it to build, train, deploy and monitor vision models for work such as defect detection, safety monitoring, inventory tracking and infrastructure inspection across manufacturing, logistics, healthcare, energy and robotics.
The platform covers the full path from data to production. Users can run inference on more than 50 supported models, train models across dozens of architectures, and build Workflows that combine models, logic and integrations. Roboflow Agent is positioned as a way to go from a described idea to a deployed vision solution. Models can run on Roboflow's serverless GPUs or be self hosted on NVIDIA Jetson, Raspberry Pi or a GPU server using the open Inference package. Other features include Vision Events, which turns detections into queryable data, Active Learning, and model weight downloads. Workspaces and models are private by default, with optional sharing to Roboflow Universe.
Pricing is credit based. The Free tier costs $0 and includes 10 credits a month. The Core plan starts at $39 per month billed monthly for 20 credits in total, with larger monthly credit packs available up to $1,399 per month for 500 credits. Enterprise pricing is custom and adds priority GPU access, volume pricing, an uptime SLA, US or EU data sovereignty and enterprise support. Serverless inference is billed per image in credits, which varies by model.