Baskerville
The Bot Hound
AI-powered protection against bots, AI agents, scrapers, and DDoS.
Clean your traffic without disrupting real users.

99.95%
Bot identification accuracy
50M+
Daily records processed & predictions returned
45+
Behavioural features modelled
100% 
Made in Canada
Key features
Built to Stop Evolving Threats

DDoS Mitigation
Baskerville reacts to anomalous network behaviour, detecting algorithmic (bot) activity and preventing malicious traffic from reaching its destination. Suspicious activity is challenged and eventually blocked.

Bot Management
Baskerville can be deployed inside any enterprise network for advanced bot management. Filter bots by activity, provenance, destination and behavioural patterns. Limit AI agents, web scrapers and other automated traffic you do not want.

AI Firewall
Automatically generate dynamic firewall policies that keep your applications secure. With an integrated AI assistant, you can ask questions, create, and manage protection policies in natural language—making enterprise-grade security as intuitive as a conversation.
Smart DDoS Mitigation
AI-powered defense that separates real users from malicious traffic in real time.
Baskerville defends against DDoS attacks at the application layer (Layer 7) by analyzing raw weblogs as input. We’ve trained Baskerville to recognize what legitimate traffic on the Internet looks like, and how to distinguish it from malicious requests attempting to disrupt your digital platforms. Baskerville has turned out to be very handy for mitigating DDoS attacks, and for correctly classifying other types of malicious behaviour.

Bot management
Deploy Baskerville for insight and control of automated web traffic.
Baskerville learns what normal internet traffic looks like and is able to detect when bots visit your platform. Baskerville gives you full visibility and management by classifying bots into categories such as: search engine crawlers, scrapers harvesting your content, AI bots consuming your content for model training, and malicious bots launching attacks or probing your platform for weaknesses.

AI Firewall
Active and adaptive defenses using AI to detect AI — leave them to sort it out!
Baskerville can be deployed in parallel or as a replacement to your web application firewall. Unlike static rate-limiting, Baskerville adapts to evolving threats and accurately blocks malicious traffic while keeping legitimate users safe.
Behind the scenes, Baskerville processes incoming web logs from Kafka, raw files, or Elasticsearch. It groups requests by host and IP, extracts behavioral features, and predicts whether they’re malicious using machine learning trained on real-world data. Results are saved to a Postgres database, with metrics available in Prometheus or Grafana for monitoring.

deployment
Works seamlessly across platforms & networks

Baskerville powers the Deflect network, protecting websites from bot attacks in real time. It analyzes traffic patterns and challenges suspicious activity before it causes harm.

Seamlessly integrated with Cloudflare Workers, Baskerville processes traffic at the edge, providing WAF-like protection, analytics, and full operator control.

COMING SOON
An official AWS Marketplace app will make it easy to deploy Baskerville inside your cloud environment.

COMING SOON
A simple plugin to bring AI-powered bot defense to WordPress sites.
How Baskerville Powers Deflect
Protecting websites from malicious bots in real-time with open-source resilience.
53,315
Challenged (past 24 hours)
99%
Precision (past 24 hours)
65
Passed Challenge (past 24 hours)
Baskerville is a machine operating on the Deflect network, protecting our clients from malicious bots. Baskerville responds to web traffic, analyzing over 50M daily requests in real-time, challenging those that are acting suspiciously.
TECHNOLOGY
Detecting anomalies in real-time, beyond traditional methods

Baskerville is built on an unsupervised anomaly detection algorithm, Isolation Forest, which does not require a labelled dataset for training. We improve on the original algorithm in order to support not only numerical but also string features from the exhibited behaviour itself.
Contact
