Demos

Things easier to show than to describe. Three of these are the real Go tool compiled to WebAssembly and running in your browser — not a reimplementation of it. The ranger demo is different: it is a report the Go tool generated, with the incident data embedded. Nothing is uploaded.

• ranger — find the service that actually broke root-cause localization
  A real incident from the OpenTelemetry Demo, localized to one service out of 87 ranked operations. Duration says every service on the request path is guilty; self time says which one did the work. This one is a report ranger wrote, not the tool running live — every trace is embedded in the page, and nothing is fetched.
• paladin — see what your agent reads integrity monitoring
  Paste a CLAUDE.md, .cursorrules or any instruction file and watch the characters your font refuses to draw appear: zero-width spaces, bidi overrides that reorder text on screen, and Cyrillic letters wearing Latin faces. It runs paladin itself, compiled to WebAssembly — not a reimplementation.
• warden — what a decline is allowed to say rules engines
  A rules engine answers in its own vocabulary — rule ids, working-memory facts, the margin floor that priced the decision. Switch between a public, partner and internal caller and watch what crosses the boundary change, with everything held back listed underneath. It runs warden itself, compiled to WebAssembly.
• genai-interlingua — normalize a GenAI span OpenTelemetry · three pages
  Six GenAI dialects spell the same token count six different ways, so nothing about your model calls can be queried as one thing. Pick one of nine real captures — they collapse onto six dialects, because LangChain's spans turn out to be OpenLLMetry's — or paste your own span, and see it rewritten into one gen_ai.* schema with what the translation cost recorded on the result. Then two pages that go further: the inspector takes one attribute at a time and says which source key produced it, what evidence the reading turned on, and what it could not carry; and what translation costs draws the measurements across all six dialects — the scoring race that picks a dialect, and where 105 attributes with no home in the conventions end up. Where it goes in your pipeline draws the three places it can run, including the one where it emits config for a Collector that has never heard of it. How it works, in pseudocode.

Not built yet

A public, live view of LLM gateway traffic: which team, which model, what was redacted, and what it cost. The normalization underneath it exists; the view does not.