EyePop.ai is a computer vision platform that lets developers and businesses build, train, and deploy custom visual AI models without deep machine learning expertise. It offers pre-trained models (person detection, object detection, pose estimation, OCR, vehicle detection) and self-service training for custom models, with deployment options across cloud, on-premise, and edge devices (NVIDIA Jetson, Qualcomm Snapdragon). The platform is SDK-first (Python and Node/TypeScript) with a visual Workflow Designer also available, and powers use cases including surveillance, insurance damage assessment, retail analytics, and broadcast highlight generation. It was founded in 2023 by Brad Chisum (CEO), Torsten Schulz (CTO), and Andy Ballester (CPO, GoFundMe co-founder), and raised a $2.85M seed round led by Innosphere Fund in February 2025. Pricing: Cloud Production at $200/mo (includes 4,000 compute units), Enterprise at $800+/mo, and On-Premise per-box licensing starting at $250/mo (1–10 boxes) with client-purchased hardware (~$1,500/box) and a one-time setup fee starting at $5,000. Critical caveats: the platform is SDK-first, not a simple point-and-click tool despite marketing language suggesting otherwise; independent user reviews are extremely limited; and initial training for custom models still requires cloud connectivity even for on-premise deployments.

Pros & Cons:

Pros: The founder team brings credibility: Brad Chisum previously co-founded Lumedyne Technologies (acquired by Google in 2014), Torsten Schulz has 20+ years scaling startups including Yahoo!, and Andy Ballester co-founded GoFundMe . The platform was recognized at ISC West 2026 with two SIA New Product awards (Judges’ Choice and Video Analytics), selected by 38 expert judges . Deployment flexibility is genuine: the same pipeline definition runs unchanged across cloud, on-premise, and edge devices including NVIDIA Jetson and Qualcomm Snapdragon, which is valuable for surveillance and industrial use cases where data cannot leave the local network . Pre-trained models cover common needs (object detection, person/pose, OCR, vehicle detection, license plates) and are available to every account without setup . The platform is HIPAA certified, and the company states data is never repurposed or shared outside the user’s specified use case . A free sandbox license for one internal lab box (3 months) allows on-premise validation before committing to hardware .

Cons: The marketing describes the platform as “no-code” and suitable for non-technical users, but the primary interface is the Python/Node SDK; a visual Workflow Designer exists, but the company itself states it is “built SDK-first for software developers and AI coding agents, not as a point-and-click tool” . Independent user validation is extremely sparse: G2 and Capterra listings exist but show no substantive reviews or ratings for EyePop.ai specifically ; RightAIChoice scanned public sources in August 2026 and could not confirm that the discussions it found were about this product rather than something else sharing its name, so it published no sentiment score . Pricing is opaque at the entry level: the $200/mo Production tier is real, but the free tier’s exact limits are not clearly documented, and on-premise requires purchasing hardware plus a $5,000+ setup fee . Custom model training has a cloud dependency even for on-premise deployments, meaning a fully air-gapped workflow is not available . The product’s positioning has shifted significantly since its 2023 founding (from general self-service CV to SDK-first developer tooling and Physical AI), which may create uncertainty about roadmap priorities.

Pricing: Cloud: Production $200/mo (includes 4,000 compute units; overage $0.05/unit; transient session minutes $0.02/min). Enterprise: $800+/mo (dedicated servers, unlimited usage, always-up servers, custom SLA). On-Premise: per-box licensing $250/mo (1–10 boxes), $200/mo (11–100), $150/mo (101–1,000); client purchases hardware (~$1,500/box estimated); one-time setup and model tuning starting at $5,000; free 3-month lab box license for testing. Signature onboarding starts at $5,000. Training iterations capped at 25/month on cloud tiers. Prices exclude any applicable taxes; exact free tier limits are not clearly documented on public pricing pages .

Who is it for? Development agencies and software teams that want to add computer vision capabilities to client projects without hiring ML engineers, using the SDK to build custom detection pipelines . Surveillance and security integrators that need on-premise or edge deployment to keep footage local while enabling real-time event detection and searchable video . Operations teams in manufacturing, insurance, retail, or agriculture that need to automate visual inspection tasks (defect detection, damage assessment, shelf monitoring) with custom-trained models . Broadcast and media teams interested in auto-clipping highlight reels from long-form footage . Not suited for non-technical users seeking a simple drag-and-drop tool without coding, individual hobbyists or solo builders (cloud starts at $200/mo, on-prem setup at $5,000+), anyone requiring a fully air-gapped solution with no internet dependency, or buyers who need extensive independent third-party reviews before committing.