eBPF Monitoring
Stream syscall activity from Docker workloads and Linux hosts with low-overhead kernel visibility.
Real-time threat detection using eBPF syscall monitoring. Zero-overhead security that scales with your containers.
Install in one line
Ship runtime protection in minutes with the official installer.
curl -fsSL https://raw.githubusercontent.com/trydirect/stackdog/main/install.sh | sudo bashBuilt for operators
Start the API with stackdog serve and stream live events through /ws.
AI on your terms
Pair local Ollama with OpenAI-compatible APIs or native Candle-backed models.
<5%
CPU overhead
25+
Built-in detectors
4
Alert channels
Real-time
Detection loop
Stackdog combines runtime telemetry, AI-assisted analysis, and automated enforcement so small platform teams can investigate and contain threats without stitching together half a dozen tools.
Stream syscall activity from Docker workloads and Linux hosts with low-overhead kernel visibility.
An Isolation Forest model learns your normal log patterns and flags behavioral drift, alongside OpenAI, Ollama, and Candle-backed analysis.
The analyzer can call back into Stackdog mid-investigation to check IP reputation, inspect container posture, review recent alerts, and ban attackers.
Discover logs, summarize incidents, archive evidence, and keep operators focused on what matters.
Ban attacking IPs through nftables or iptables, quarantine suspicious containers, and release bans automatically when the ban window expires.
Track file integrity against SQLite baselines, audit SSH, sudoers, and Docker daemon settings, and flag privileged containers or exposed Docker sockets.
Route incidents to Slack, email, or webhooks, filtered by minimum severity and deduplicated so a standing misconfiguration does not page you all day.
Prioritize with built-in signatures, heuristics, and scoring that helps teams act on the highest-risk events first.
Deploy Stackdog with a one-line install, a pinned release, or the published Docker image.
Collect syscalls, container telemetry, and logs while the rule engine and AI analyzers watch for anomalies.
Push alerts, score threats, and automatically firewall or quarantine suspicious workloads.
Stackdog ships with a clean REST API and WebSocket feed, so you can wire dashboards, runbooks, ticket automation, or chatops around the same event stream that powers the built-in security engine.
These common questions are optimized for both human readers and AI search engines, so buyers and assistants can quickly understand where Stackdog fits.
Stackdog Security is a Rust-based runtime security platform for Docker containers and Linux servers. It combines eBPF monitoring, AI-assisted log analysis, threat scoring, and automated response in one stack.
Stackdog uses eBPF-based syscall monitoring and container-aware telemetry to observe workload behavior with low overhead. It also inspects logs and Docker metadata to enrich detections.
Stackdog can work with OpenAI-compatible APIs, local Ollama deployments, and Candle-powered Rust inference workflows. Teams can choose the provider that best matches their security and cost requirements.
Stackdog can already recognize Kubernetes-style container identifiers in runtime paths, but broader Kubernetes support is still on the roadmap. Today the primary deployment target is Docker containers and Linux servers.
Stackdog can score events, notify responders, push Slack or webhook alerts, update nftables or iptables rules, and quarantine suspicious containers as part of an automated response pipeline. Bans expire on their own, and you can also ban an address by hand with stackdog ban-ip.
Alerts are filtered by a minimum severity you set, and repeats of the same finding are suppressed for a configurable window (six hours by default). Standing misconfigurations are re-detected on every pass, so without this they would notify all day. Each alert also names the real source, such as the container or log file it came from.
Yes. Stackdog is open source on GitHub under the MIT license, so teams can audit the code, run it themselves, and contribute improvements.
Stackdog is built for Docker containers and Linux servers. The core platform is written in Rust, and optional machine learning features use Candle for native Rust inference.
Install Stackdog with the one-line script or Docker image, then review the CLI and API docs to start monitoring your infrastructure. From there you can configure AI providers, alert channels, and response actions.
Install Stackdog, connect your alerting channels, and bring AI-assisted runtime detection to your Docker and Linux workloads.