Israeli AI startup Conntour raises $7 million Seed round to transform video surveillance
The platform enables security teams to search video using natural language instead of predefined rules.
Conntour, a company developing an artificial intelligence platform for real-time video intelligence analysis, has raised $7 million in Seed funding. The round was led by General Catalyst, with participation from Y Combinator, SV Angel, and Liquid 2 Ventures, along with other investors.
Conntour enables security teams to query cameras using natural language, finding any object, person, or situation without relying on preset categories or pre-programmed rules, for example, users could search for “a man with a tattoo on his left arm” or “a van with a print of fruits on it.”
The company was founded in 2024 by Matan Goldner and Tomer Kola, computer vision experts with experience in video analysis and technology companies. Conntour employs 14 people at its offices in Tel Aviv and participated in the first cycle of Palantir’s Startup Fellowship program. The idea for the company originated from working with IDF field observers during their reserve service after October 7, when both founders were in active combat reserve units.
"Traditional video surveillance forces operators to define exactly what they're looking for before they even know what they need to find," said Conntour CEO Matan Goldner. "Existing solutions can only detect a predefined set of parameters, such as a weapon or a make of car. But what do you do when you need to identify someone passing a bag to another person, or a man with a Nike shirt? Real-world security doesn't work in neat categories. Our platform brings search-engine-level intelligence to any camera network, so security teams can respond to threats as they unfold and investigate incidents in minutes rather than days.”
Conntour’s platform is designed to address that constraint by applying computer vision algorithms capable of interpreting complex, context-driven queries. The system can operate both in real-time, flagging potential threats as they occur, and retrospectively, allowing investigators to search through large volumes of archived footage.
The company says the technology is already being deployed in homeland security operations in Singapore, an early indication of its intended use in high-stakes environments such as border control, critical infrastructure, and large public venues.
The company claims its approach can significantly reduce the time and labor involved in monitoring and reviewing footage. Among the metrics it cites: the ability for a single operator to monitor thousands of cameras simultaneously, analyze large volumes of recorded video in minutes, and reduce both false alarms and missed events. Such improvements, if borne out in practice, could alter the economics of large-scale surveillance, where manual review remains costly and time-intensive.
