Problem

Conntour addresses the surveillance industry’s scale and flexibility gap: organizations may have hundreds or thousands of cameras but not enough human operators to analyze the resulting footage. Traditional video analytics rely on a finite set of predefined parameters—such as license plates, clothing color, weapons, or bags—so they struggle with context-heavy events, producing missed incidents, false alarms, and hours of manual review. At large deployments, continuously applying heavyweight AI also creates major GPU, power, bandwidth, latency, and infrastructure costs.

The killer use case is rapidly finding or detecting something that could not have been cleanly specified in advance. An investigator can search for a particular person, vehicle, or interaction across recorded footage, while an operator at a sensitive facility can receive an alert when a sequence such as someone leaving a bag and walking away occurs. Conntour’s own example is locating a distinctive van across the camera network in seconds instead of asking staff to watch thousands of feeds.

Product / Service

Conntour is an AI video-intelligence layer that lets security personnel query live and recorded camera feeds in natural language. Users can ask questions such as where a person appeared, configure plain-English real-time alerts, and extract structured information from footage. The system is designed to understand more flexible descriptions and behaviors than fixed-rule analytics, and to make it possible for one human to monitor very large camera networks.

The product can run fully on premises, in the cloud, or in a hybrid configuration, and it connects to existing video-management and command-center systems rather than requiring a new camera installation. Its architecture selectively applies computationally intensive models so it can scale more efficiently; Conntour says it can monitor up to 50 feeds from a single consumer GPU. That combination of natural-language flexibility, deployment control, and reuse of existing infrastructure is intended to turn passive recordings into searchable, actionable intelligence while preserving control of sensitive footage.

Market

Conntour competes in AI video intelligence, video-content analytics, and enterprise physical-security software, with an especially strong focus on mission-critical government, defense, law-enforcement, transportation, and enterprise environments. Its alternatives include established video-analytics providers such as BriefCam, which offers video search, real-time alerting, and quantitative insights, as well as newer AI physical-security platforms such as Rhombus and Coram. Conntour’s differentiation is its emphasis on open-ended natural-language queries across large, heterogeneous camera networks and fully on-premises deployment, rather than only predefined detections or a bundled camera system.

The company is not merely pre-revenue: Y Combinator’s company profile says it secured Singapore’s government as a paying customer within months of launch, and TechCrunch reported several large government and publicly listed customers, including Singapore’s Central Narcotics Bureau. Conntour was founded in 2024, entered YC’s Winter 2025 batch, and raised a $7 million seed round in March 2026 from General Catalyst, Y Combinator, SV Angel, and Liquid 2 Ventures. Public revenue figures and a precise customer count were not disclosed in the evidence, but the reported paying deployments and institutional funding indicate early commercial traction.