Boldstreet Logo
Boldstreet Logo

RAVEN — The Surveillance Engine for Safer Cities & Critical Sites

Protect people, secure assets, reduce response times — request a live demo of RAVEN today.

Request RAVEN DemoExplore All Products
What RAVEN Does
RAVEN is Boldstreet’s enterprise surveillance engine built for African cities and mission-critical sites. Using edge-optimized computer vision, audio analysis, thermal imaging, and sensor fusion, RAVEN detects and classifies 1,450+ anomaly types spanning human behavior, objects, vehicles, events, and environmental hazards. Designed for malls, airports, transport hubs, commercial campuses, and municipal control rooms, RAVEN turns scattered signals into verified, actionable intelligence, with evidence hashed on a blockchain ledger for auditability and legal defensibility.
  • Reduce average incident detection-to-response time by 60% (pilot estimate).
  • Cut false positives by prioritizing corroborated multi-sensor evidence.
  • Maintain chain-of-evidence using blockchain anchoring for every alert.
  • Integrate with existing CCTV, PSIMs, and emergency response workflows.
    • Use Cases — Where RAVEN Makes a Difference
    Retail & Shopping Malls

    Retail & Shopping Malls

    Detect abnormal crowd formation, unattended baggage, aggressive behavior, and access breaches to high-security back areas. Alerts mall control rooms before incidents escalate and correlates CCTV clips with incident metadata.

    Airports & Transport Hubs

    Airports & Transport Hubs

    Identify erratic passenger flows, queue surges, slip-and-fall events, and unauthorized tarmac access. Thermal and audio sensors flag sudden heat signatures or loud disturbances.

    Critical Infrastructure & Utilities

    Critical Infrastructure & Utilities

    Monitor perimeters for tampering, garbage fires, and unauthorized vehicle approaches. Blockchain-locked evidence reduces audit friction for operators and insurers.

    Municipal Security / Events

    Municipal Security / Events

    Live crowd monitoring during public events, automatic crowd-density alerts, and predictive forecasts to reposition patrols and emergency services.

    Corporate Campuses & Warehouses

    Corporate Campuses & Warehouses

    Detect tailgating at secure doors, unsafe lifting/handling, and hazardous chemical spills via integrated environmental sensors.

    Technology Stack — How RAVEN Sees & Decides

    Edge Vision (tuned models)

    • Custom vision training sets tuned to local clothing, vehicle types (including boda-bodas), and signage common in Kenya and East Africa.
    • Model suites include human pose estimation, face-blur for privacy, gaze approximation, and object detection (bags, weapons, fire sources).
    • Edge inference keeps raw video on-premise; only anonymized metadata and flagged clips are transmitted.

    Multi-Modal Fusion

    • Audio analytics for gunshot, explosion, glass-break, and aggressive shouting signatures.
    • Thermal cameras for low-light or obscured detection (smoke, heat from vehicles or fires).
    • Environmental sensors (CO2, gas, temperature) for hazardous-event correlation.

    Streaming & Processing

    • Kafka-based streaming ingestion with micro-batch and real-time pipelines.
    • Distributed microservices for rule-based correlation, anomaly scoring, and risk prioritization.

    Evidence Integrity

    • Each verified alert includes a hashed evidence bundle recorded on a permissioned blockchain.
    • Ensures tamper-proof audit trails for regulators, insurers, or law enforcement.

    Interoperability

    • Open REST APIs, ONVIF support for cameras, and connectors for PSIM and existing NVR systems.
    • Webhooks, SMS, email, and secure control-room integrations for alert routing.
    Detection Coverage & Metrics

    Human-behavior anomalies

    • Loitering > threshold time
    • Aggressive gestures / fighting
    • Sudden directional surges (crowd panic)
    • Tailgating at access points

    Object & item anomalies

    • Unattended baggage (left-object detection)
    • Weapon detection (shapes and silhouettes)
    • Hazardous object (smoke, fire-carrying materials)

    Vehicle & perimeter anomalies

    • Abandoned vehicle detection
    • Wrong-way / counterflow
    • Illegal parking in restricted zones

    Environmental & event

    • Fire/heat spikes, gas leaks (via sensor fusion)
    • Glass-break audio signature
    • Crowd density and dwell heatmaps (people/m²)

    Performance / accuracy metrics

    • Model precision / recall dashboards (95%+ on validated datasets)
    • False positive rate reduction via multi-sensor correlation
    • Mean Time To Detect (MTTD) & Mean Time To Respond (MTTR) reporting
    How RAVEN Fits Your Operations

    Phase 0 — Risk & Systems Assessment

    1–2 weeks

    Asset discovery, camera audit, comms review, risk scoring, and SLA alignment. Deliverable: Implementation blueprint + ROI projection.

    Phase 1 — Pilot Deployment

    4–8 weeks

    Edge boxes & camera configuration for 3–10 sites, integrate 1–2 sensor types, dashboard access, and on-site training.

    Phase 2 — Scale & Harden

    4–12 weeks

    Full rollout, integration with PSIM/NOC, multi-site orchestration, dedicated incident rules and playbooks.

    Phase 3 — Operate & Optimize

    Ongoing

    24/7 support, model retraining with site-specific data, monthly tuning for new event types and seasonal patterns.

    Turnkey services available: hardware procurement, installation, site surveys, training, and runbooks for security teams. We also offer managed SOC services for customers who want a fully outsourced monitoring option.
    Privacy, Compliance & Security
    RAVEN is built with privacy-by-design: on-edge anonymization (faces and PII blurred unless lawfully overridden), data minimization (only metadata and short clip extracts stored), GDPR and local compliance (configurable retention policies and access logs), and secure infrastructure (TLS, AES-256, role-based access control). Blockchain-anchored evidence ensures immutable audit trails, reducing legal risk and strengthening law enforcement relationships.
    Competitive Differentiation
    Localization & Context: Models trained on African datasets for higher real-world accuracy.
    Sensor Fusion for Confidence: Audio, thermal, and environmental sensors reduce false positives.
    Evidence Chain: Blockchain hashing for audit-grade integrity not offered by standard VMS/NVR vendors.
    Platform Unification: Interoperates with Boldstreet engines for city planning or dynamic ad bidding.
    Pilot-to-Scale Path: Designed for small pilots that scale to city-wide deployments.
    Pricing & Pilots

    Pilot — Secure Start

    3 sites, 6-week pilot, dashboard access, 24/7 alerting, training. Price: $5,000 – includes hardware leasing options.

    Scale — Enterprise

    Multi-site, API integrations, managed SOC. Pricing: custom, SLA-backed.

    Exact pricing depends on site complexity, number of cameras, sensor types, and integration needs. Contact sales for a scoped quote.
    Case Snapshot

    Objective

    Reduce security blind spots, test unattended-baggage & crowd detection.

    Approach

    Deployed RAVEN across 6 camera zones, integrated audio sensors, and configured playbooks with mall control.

    Outcome

    RAVEN detected simulated unattended items and crowd surges in under 45 seconds; patrol response time reduced by 54%. Mall management reported fewer false alerts and faster incident resolution. Full evidence bundles were blockchain-hashed for audit and insurance purposes.

    Two Rivers Mall case study
    FAQs

    Next Steps — Book a Live Demo
    RAVEN is ready for pilots across Nairobi and East Africa. Schedule a demonstration to see live detection, end-to-end evidence capture, and how RAVEN integrates with your teams. Download our detailed proposal to learn more.
    Request RAVEN DemoDownload Proposal