Conntour secures $7M from General Catalyst and Y Combinator to develop an AI-powered search for security video

The surveillance technology industry is currently under intense scrutiny, though not necessarily for positive reasons. Ongoing controversy surrounds agencies such as U.S. Immigration and Customs Enforcement using networks such as Flock to monitor individuals. At the same time, home security company Ring has faced criticism for introducing features that could allow law enforcement to request neighbourhood footage from homeowners. These developments have sparked broader debates around safety, privacy, and control over surveillance.

Despite the controversy, market demand remains strong, especially as advances in vision-language models continue to accelerate innovation in monitoring technologies. Companies are increasingly exploring new ways to observe and analyse activity within physical spaces.

According to Matan Goldner, co-founder and CEO of Conntour, ethical considerations are central to the company’s strategy. He emphasised that Conntour is selective about its customers, even though it is still a relatively young startup. Goldner explained that the company can afford to be selective because it has already secured several major clients, including government organisations and publicly listed companies, such as Singapore’s Central Narcotics Bureau.

“The fact that we have such big customers allows us to select them and to stay in control […] We’re really in control of who is using it, what is the use case, and we can select what we think is moral and, of course, legal. We use all our judgment, and we make decisions based on specific customers that we’re okay [to work with] because we know how they will use it,” Goldner said in an interview.

This early traction has also attracted investor interest. Conntour recently raised $7 million in seed funding from General Catalyst, Y Combinator, SV Angel, and Liquid 2 Ventures.

Goldner noted that the funding round was completed extremely quickly. “I think I scheduled around 90 meetings in like eight days, and just after three days — we started on Monday and by Wednesday afternoon, we were done,” he said.

Conntour’s cautious approach may be justified given the capabilities of modern AI-powered surveillance tools. Its platform uses AI models to enable security teams to search camera feeds with natural language queries, effectively serving as a Google-like search engine for video footage. The system can identify objects, individuals, or specific scenarios in real time, autonomously detect threats based on predefined rules, and generate alerts.

Unlike traditional surveillance systems that rely on fixed parameters or predefined rules to detect motion or behaviour, Conntour’s solution leverages vision-language models, offering greater flexibility and ease of use. For example, a user could query, “Find instances of someone in sneakers passing a bag in the lobby,” and the system would scan both recorded and live footage to return relevant results.

The platform also integrates AI capabilities that allow users to ask questions about video content and receive text-based responses, accompanied by relevant clips. Additionally, it can automatically generate incident reports based on identified events.

One of Conntour’s key differentiators is its scalability. Goldner explained that the platform is designed to handle thousands of camera feeds efficiently. In some cases, it can monitor up to 50 camera streams using a single consumer-grade GPU, such as Nvidia RTX 4090.

This efficiency is achieved through a system that dynamically selects the most appropriate models and logic pathways for each query, minimising computational load while maintaining performance.

The platform is also flexible in deployment. It can be implemented entirely on-premises, fully in the cloud, or in a hybrid setup. It integrates with existing security infrastructure or can function as a standalone surveillance solution.

However, the industry continues to face a fundamental limitation: the quality of analysis is constrained by the quality of the footage itself. Poor lighting, low-resolution cameras, or obstructed lenses can significantly reduce accuracy.

To address this, Conntour provides a confidence score alongside search results, allowing users to gauge the reliability of the output. Lower-quality footage leads to lower confidence levels, helping users interpret the results more effectively.

Looking ahead, Goldner identified a key technical challenge: balancing advanced natural language capabilities with system efficiency.

“We have two things that we want to do at the same time, and they contradict each other. On the one hand, we want to provide full natural-language flexibility, LLM-style, so that you can ask anything. And on the other hand, there’s efficiency, so we want to make it use very few resources, because again, processing [thousands] of feeds is just insane. This contradiction is the biggest technical barrier and technical problem in our space, and what we’re working really, really hard to solve.”