Conntour, a startup developing AI-powered search for video surveillance systems, has raised $7 million in a seed funding round backed by General Catalyst, Y Combinator, SV Angel, and Liquid 2 Ventures.
The round highlights growing investor interest in AI-driven security infrastructure, as advances in vision-language models unlock new ways to analyse and interpret video data at scale.
Turning Surveillance Into Search
Conntour is building what it describes as a “search engine for security footage.” Its platform allows users to query live or recorded video feeds using natural language — similar to how one would search on Google.
Security teams can ask questions like identifying specific behaviors, objects, or interactions, and the system returns relevant video clips alongside text-based explanations.
Unlike traditional surveillance tools that rely on predefined rules, Conntour leverages advanced vision-language AI models to deliver more flexible and context-aware results.
Advertisement
Real-Time Monitoring And Automated Alerts
Beyond search, the platform can continuously monitor video feeds and detect potential threats based on predefined conditions.
It automatically surfaces alerts and generates incident reports, helping organisations respond faster while reducing manual monitoring efforts.
Built For Scale
One of Conntour’s key differentiators is its ability to scale across large surveillance networks.
The system is designed to handle thousands of camera feeds efficiently, with the ability to process up to 50 streams on a single consumer-grade GPU such as the Nvidia RTX 4090.
To achieve this, Conntour dynamically selects the most efficient AI models for each query, balancing performance with computational cost.
Flexible Deployment Across Environments
The platform can be deployed fully on-premise, in the cloud, or in hybrid environments, and integrates with existing security systems or operates as a standalone solution.
To address limitations in video quality — a common issue in surveillance — the system provides confidence scores for its outputs, helping users assess the reliability of results.
Balancing Power And Efficiency
According to co-founder and CEO Matan Goldner, the biggest technical challenge lies in combining full natural language flexibility with computational efficiency.
Delivering powerful, LLM-like querying across thousands of video feeds requires significant optimisation to ensure the system remains scalable and cost-effective.
Ethical Considerations And Selective Customers
As debates around surveillance, privacy, and data use intensify, Conntour is taking a selective approach to its customer base.
The company works with large enterprise and government clients, including Singapore’s Central Narcotics Bureau, and emphasises careful consideration of use cases and ethical implications.
About Conntour
Conntour is an AI-driven video intelligence platform that transforms surveillance systems into searchable, automated insights engines. By combining vision-language models with scalable infrastructure, the company enables organisations to monitor, analyse, and respond to real-world events in real time.
