Documentation

RunLore investigates incidents on your platform and remembers what it learns.

New here? Getting Started runs a real investigation on recorded evidence in about a minute — no cluster, no API key — then walks up to a production install.

Where to go next depends on what you need:

• Integrations — wire RunLore to what you already run: what triggers an investigation, which LLM runs it, the data sources it reads, where notifications land, and the Git forge it writes knowledge back to.
• Concepts — how it is designed and why it gets better over time, rather than how to configure it.
• Configuration — the exhaustive key reference every other page deep-links into.
• Operations — running it: metrics, logs, diagnosing “why didn’t it do X?”, and upgrades.
• Security — what an LLM-driven agent pointed at your cluster can and cannot do, and what enforces that.
• Reference — the lore CLI, model benchmarking, the kb-steward skill, and worked examples.