XERJ · Product Tour

~90 seconds · one 22 MB binary, no JVM · Elasticsearch-compatible API · live AI-telemetry dashboards

↓ Download MP4 (8 MB) Press space to play/pause; ←/→ to seek 5s; click the timeline to scrub.

What the tour shows

A single-take walkthrough of XERJ — one self-contained binary that speaks the Elasticsearch REST API and ships its typography-first console inside the same executable (no JVM, no separate Kibana). The tour opens on live operational dashboards built for AI/LLM workloads and moves through the console's sections:

AI · OVERVIEW — query volume, token spend, and cost broken down by model.
RAG · QUALITY — grounding score, hallucination rate, retrieval hit rate, and citation counts.
VECTOR · INDEX — vector search throughput and recall.
AGENT · MEMORY — stored entries, dedup rate, recall, and per-agent activity.
DATA / DISCOVER / ALERTS / SETTINGS — the indices behind the dashboards, an interactive query console, alert destinations, and role/passkey settings.

Every number on screen is rendered from documents posted to XERJ over its /_bulk and /_search endpoints — the same API an Elasticsearch client already uses. Prefer the short loop? The animated preview on the home page is a clip of this same tour.