Scylla
Creating the best possible AI tech for video surveillance
Est 2018|Valuation $220K|Raised $5.7M
MarketsPhysical Security Systems
Creating the best possible AI tech for video surveillance
Published Oct 5, 2026
The vendor references low-latency edge inference and mentions sub-100ms per frame latency for Event Raptor in its announcement, but real-world performance will vary by:
Given limited Scylla-specific churn evidence in reviews, likely churn reasons in this category include:
Common selection drivers represented in vendor materials:
A February 2026 press/partner write-up describes Scylla partnering with a large government-focused reseller to sell into U.S. public sector contracting vehicles (SEWP V, NASPO ValuePoint, etc.), and repeats vendor-reported scale claims.
Critical evaluation: Availability through procurement channels can reduce friction for government buyers; however, it is not itself proof of efficacy.
Claims such as large frame-processing figures and asserted false-alarm reductions are not accompanied by transparent definitions or an audit trail in public sources.
Based on per-camera pricing and monitoring-center orientation, Scylla appears suitable for:
Get a comprehensive analysis of Scylla including market position, competitive landscape, adoption trends, and peer benchmarks.
Download full report →From the very limited G2 review coverage (the reviewer disclosed a vendor relationship): the reviewer reported that Scylla detected guns with apparent high reliability and low false positives, including detections in reflections on glass surfaces. The review also noted that optimal results required correct camera placement and configuration.
Interpretation: While vendor-facing reviews offer some positive signals, the public sample is very small and the disclosed relationship reduces the evidentiary weight.
Because Scylla AI has limited direct end-user review coverage on major public platforms, the most accessible skeptical sources relate to the broader category of AI gun detection rather than Scylla specifically. These discussions highlight risks such as operational uptime, integration maturity, and reputational or governance concerns associated with surveillance deployments.
Practical takeaway: For Scylla specifically, the public review footprint is thin; procurement teams should compensate via pilots, reference calls, and contract guardrails.
Scylla publishes multiple partnership announcements and integration news:
Critical evaluation: These are primarily vendor-authored announcements. They indicate integration breadth but generally do not provide independently verified outcomes (e.g., before/after alert volumes, measured response time reduction, precision/recall in the customer’s environment).
Scylla (Scylla AI / Scylla Technologies Inc.) is an AI video analytics platform for physical security. It analyzes live video streams (and recorded footage in some offerings) to detect threats and anomalies such as firearms, perimeter intrusion, fighting/aggression, abandoned objects, smoke/fire, and other safety/security events, then routes alerts (with context such as screenshots, time, and location) to operators and response workflows. The vendor presents the platform as intended to reduce false alarms and improve response speed by adding an additional analytics layer to existing camera/VMS deployments.
Key architecture themes in Scylla’s materials include edge/on-prem deployment options (including air-gapped environments), integrations with VMS platforms, and modules packaged for different detection problems. Scylla offers a computer-vision backbone referred to as ScyllaNet and a decision layer referred to as Charon.
Scylla’s messaging and partner ecosystem indicate a primary focus on:
It competes in the AI video analytics / weapon detection / threat detection segment—often as an add-on to existing camera and VMS infrastructure.