Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Today the surveillance technology industry is in the spotlight, but not for the best reasons. With the controversy surrounding US Immigration and Customs Enforcement Eavesdropping on the Flock camera network To monitor people, home camera maker Ring Cash fee To build new features that would enable law enforcement to ask homeowners for footage of their neighborhoods, there is currently widespread debate about safety, privacy, and who can watch whom.
But controversy doesn’t erase markets, and the continued improvement of vision language models has only put more wind into the sails of companies working to build new ways to help companies monitor what’s going on in their buildings.
According to Matan Goldner, co-founder and CEO of the video surveillance startup Contourethics around this topic are important enough that he says his company is quite selective about which customers it sells its products to. This may not make sound business sense for a startup barely two years in, but Goldner says he can afford to do so because Contour already has several large government and publicly listed clients, one of which is Singapore’s Central Narcotics Bureau.
“The fact that we have such large clients allows us to choose them and stay in control (…) We really have control over who uses it, what the use case is, and we can choose what we think is ethical, and of course legal. We use all of our judgment, and we make decisions based on specific clients that we’re okay (to work with) because we know how they’re going to use it,” Goldner told TechCrunch in an exclusive interview.
This traction has helped Contour do more than just be selective. Investors noted that the startup recently raised a $7 million seed funding round from General Catalyst, Y Combinator, SV Angel, and Liquid 2 Ventures.
The tour closed within 72 hours, Goldner said. “I think I scheduled about 90 meetings in about eight days, and after just three days — we started on Monday and by Wednesday afternoon, we were done,” he said.
Regardless, Contour may be right in his choice, especially considering how powerful AI tools are in this area. The company’s video platform uses AI models to allow security personnel to query camera feeds using natural language to find any object, person or situation in the footage, in real-time — a Google-like search engine designed specifically for security video feeds. It can also monitor and detect threats on its own based on pre-defined rules, and surface alerts automatically.
Unlike legacy systems that rely on pre-defined definitions or parameters to detect specific objects, movement patterns, or behaviors, Contour claims that its system uses natural language and vision models, giving it a high degree of flexibility and ease of use. A user might ask: “Find instances of a person wearing sneakers passing a bag in the hallway,” and the Conntour system will quickly search through all recorded footage or live video feeds to return relevant results.
Because the platform relies on artificial intelligence models, users can simply ask questions about the footage and get text answers, accompanied by relevant video feeds, as well as generate incident reports.
However, the company’s selling point is its scalability. Goldner explained that the platform is fundamentally different from other AI-powered video search services because it is designed to scale efficiently in systems with thousands of camera feeds. In fact, he said the Conntour system can monitor up to 50 cameras from a single consumer GPU like Nvidia’s RTX 4090.
The company does this by using multiple logic models and systems, then determining which models and systems the algorithm should use for each query to require the least amount of computing power to give users the best results.
Conntour claims that its system can be deployed entirely on-premises, entirely on the cloud, or a combination of both. It can be connected to most security systems already in use, or it can serve as a complete monitoring platform on its own.
But there has been a long-standing problem in the video surveillance industry: the quality of surveillance is only as good as the quality of the captured footage. It’s difficult to tell details from poorly lit parking lot footage recorded by a low-resolution camera with a dirty lens, for example.
Goldner says Contour hedges against this inevitability by providing a confidence score alongside its research findings. If the camera feed is not of good enough quality, the system will return results with low confidence levels.
Moving forward, Goldner says the biggest technical problem to solve is bringing the full level of LLM capability to his system while maintaining its efficiency.
“We have two things we want to do at the same time, and they conflict. On the one hand, we want to provide complete flexibility in natural language, LLM-style, to allow you to 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 crazy. This contradiction is the biggest technical barrier and technical problem in our field, and what we’re really working on, it’s very difficult to solve.”