Frontend SDK
Drop into any SPA, SSR, or static site
- Next.js · Nuxt · Remix · Astro
- React · Vue · Svelte · SolidJS
- Web Vitals auto-binding · custom metric hooks
ZoneFlash is the only edge-native RUM platform that ingests user-experience telemetry at the nearest collector, then ships a single composite score to your console. Median time-to-first-insight is 792 ms — measured end-to-end, from browser event to dashboard cell, in the 2024 WebPerf Audit.
Trusted by 1,840+ engineering teams · deployed in 47 countries · 9.4T events processed in 2024
Every visitor event lands at one of 312 globally distributed collectors — never at a distant monolith. From there it travels through our QUIC-based stream aggregator and is scored against the Real Experience Score model within the same browser tick. There is no overnight batch window, no five-minute polling delay, no “explore” pane you have to refresh.
Each PoP hashes and rolls up events at the edge before they leave the region, so customer PII never traverses the public internet twice.
Encrypted, multiplexed, and back-pressured. Aggregators fan into the RXS engine without dropping sessions under load spikes.
Alerts, dashboards, and Slack/PD hooks all subscribe to the same stream — no second pipeline to maintain.
No agent zoo. No sidecar sprawl. Whether you ship a React SPA, a Cloudflare Worker, or a Rust service on bare metal, the same RXS surface shows up in your console within the first deploy.
Drop into any SPA, SSR, or static site
Instrumentation at the request boundary
Server-side telemetry, one agent per host
Pull logs from where the bytes already live
Every tile below is a first-class integration, not a community plugin. If your stack is here, the console will be live before your next deploy window.
Don’t see your stack? Our SDK is open-core and the collector speaks OTLP out of the box.
Ask about a custom integrationThe Real Experience Score (RXS) is a 0–100 weighted aggregate of four signal families: load timing (LCP and TTFB), interactivity (INP and long-task duration), visual stability (CLS), and business criticality (per-route conversion impact, supplied by your analytics export). The default weights were derived from a regression against 2.1 billion anonymized sessions across 41 referenceable customers in 2023, and are re-fit quarterly.
What makes RXS different from a single Web Vital is the criticality term. A 200 ms LCP regression on your checkout route hurts RXS far more than the same regression on a marketing landing page — the score reflects what your users actually convert on, not just what they happen to load. Every weight in the model is exposed in the whitepaper, alongside the bootstrap procedure and the false-positive rate we measured against the 2024 WebPerf Audit ground truth.
For the serious reader, the appendix covers the rolling-window decay function (we use a 28-day half-life, not a naive mean), the confidence interval we publish per segment, and the exact formula we use to attribute a regression to a specific deploy. We publish the methodology because we want the score to be defensible — it is the only metric on the dashboard your CTO will be asked about by the board.
On the call we’ll spin up a sandbox tenant against your production hostname, replay your last seven days of traffic through the RXS engine, and drill from a 312-PoP world map down to a single Core Web Vital on a single route. You’ll leave with a CSV of regressions and the median time-to-first-insight we measured on your stack.
No slide deck. No sales engineer monologue. A console, your routes, and 20 focused minutes.