Data Filters
Filter and clean sensor data with Deadband, Debounce, Discard Above, and Discard Below filters.
Krill is the first decentralized control platform where every sensor, actuator, camera, and logic rule is a node in a self-healing mesh — running on your Raspberry Pis, owned by you, with native apps on every platform.
Built from the ground up in Kotlin Multiplatform. One codebase. Android, iOS, Desktop, Web, and Raspberry Pi. No cloud. No subscription. No single point of failure.
Your data never leaves your network. No accounts, no subscriptions, no vendor lock-in.
Servers auto-discover each other on your LAN. Add a Pi, enter a 4-digit PIN, and it joins the swarm.
Logic gates, Python lambdas, serial sensors, MQTT, GPIO, cameras, color sensors, and 35+ node types — wired together visually.
A unique force-graph shows live alerts and executions — or design custom SVG dashboards in Inkscape with real-time data and live video overlays.
Everything in Krill is a Node. Nodes react to data changes, execute logic, and control hardware — across any number of servers, in real time.
Filter and clean sensor data with Deadband, Debounce, Discard Above, and Discard Below filters.
Trigger workflows on a schedule using cron expressions for time-based automation.
Generate statistical summaries of data over time periods for trend analysis.
Compute derived values from other data points using formulas and aggregations.
Manual workflow triggers activated by user interaction for on-demand automation.
Use logic gates and GPIO pins to build leak detection safety interlocks for aquarium and hydroponic water systems with Krill.
Understanding how Krill servers and clients manage TLS certificates for secure peer-to-peer communication
Deep dive into Krill's peer-to-peer mesh networking including beacon discovery, cluster PIN trust, SSE real-time updates, and server settings bootstrap
Understanding how Krill servers use API keys for client authentication and how to manage them
Understanding how Krill securely executes user-provided Python scripts with multi-level sandboxing and isolation