The most accurate analytics
you can run. Privately.
Metrone is an analytics platform engineered for precision. Every event — page view, voice call, chat session, conversion — is captured at the edge with zero data loss, zero cookies, and zero compromise on visitor privacy. Built for teams shipping AI-powered products who refuse to choose between good data and good ethics.
Where traditional analytics estimate and sample, Metrone measures. Where they require consent banners that drive away visitors, Metrone needs none. Where they ignore AI interactions entirely, Metrone tracks them natively — and gives AI agents the same programmatic access as your human team.
What makes Metrone different
First-party script loads from your domain. No ad blocker interference. Every visitor counted. The data you see is the data that exists — not a sampled, filtered, consent-gated fraction of it.
No cookies, no IP storage, no fingerprinting, no consent banners. Privacy isn't a mode you enable — it's the architecture. GDPR, CCPA, and PECR compliant by design, not by configuration.
Page views, voice calls, chat sessions, copilot queries, and conversions flow through one pipeline into one dashboard. No stitching tools. No fragmented data. One source of truth for web and AI.
Two first-class interfaces: one for humans who read dashboards, and one for AI agents that consume APIs. Same data, same privacy, different access patterns.
Platform capabilities
For Humans
One DNS record, tracking runs from your domain.
Zero-dependency, ~2 KB. Auto page views, custom events, conversions.
Provider, hooks, and HOC for React 18+ apps.
Cloudflare Workers, Vercel Edge, or Netlify Edge.
Built-in AI assistant in every dashboard — ask questions about your analytics in natural language.
Per-project timezone settings with local-time bucketing.
For AI Agents
Node.js / Deno / Bun. Auto-batching, retry, offline queue.
Model Context Protocol — AI assistants query and track analytics natively.
Twilio, VAPI, Retell, Bland, OpenAI, Anthropic, Telnyx, Intercom, Zendesk — paste a URL, done.
Full read/write. Batch query, cursor pagination, idempotency.
Server-Sent Events stream for live dashboards and monitoring.
Machine-readable error codes with request tracing.
Our principles
Accuracy over convenience
Most analytics tools trade accuracy for easy setup. Third-party scripts get blocked by 42% of users. Cookie-based tracking loses 70% of EU data. Metrone chose the harder path — first-party architecture — so your numbers actually reflect reality.
Privacy as a constraint, not a feature
We didn't add a privacy toggle. We designed the entire system so personal data never enters it. IP addresses are hashed with a rotating daily salt at the edge and discarded. The raw data doesn't exist to leak, sell, or subpoena.
AI-native from day one
The schema was built for AI from the start. Voice agent calls sit next to page views. Chatbot resolution rates sit next to conversion funnels. Server SDK, MCP protocol, real-time SSE, and batch query API give AI agents the same access as your dashboard.
Developer-first access
REST API and webhooks on every plan. CSV and JSON exports. Real-time SSE streaming. Zero-code webhook adapters for AI providers. Server SDK for Node.js, Deno, and Bun. MCP integration for Claude Desktop and Cursor. Your data is yours.
Timeline
Metrone built as internal analytics engine inside Sentrix.
Voice and AI agent tracking added (Twilio, VAPI, Retell). Metrone name born — independent product direction set.
Public beta launch. Browser SDK, React SDK, CNAME setup, webhook integrations. AI-era update: Server SDK, MCP protocol, SSE, batch query API, structured errors.
Built-in AI assistant for every project dashboard. Multi-project management, SDK v1.2.0, team invitations.
Origin
Metrone started inside Sentrix — an AI-powered digital presence management platform. The need was specific: accurate analytics that could track both web traffic and AI interactions (voice agents, chatbots, copilots) without storing personal data. Every existing tool failed on at least one of those requirements.
The internal tool became a product when it became clear the gap wasn't just ours. Every team shipping AI features faces the same problem. Metrone is now available as a standalone platform, built in Silicon Valley, for anyone who demands precision from their analytics without compromising on privacy.
Since launch, Metrone has expanded beyond web analytics into a dual-interface platform. Humans get dashboards, timezone-aware reports, and a built-in AI assistant. AI agents get a Server SDK, MCP protocol, REST API, real-time SSE, and structured errors — everything they need to read and write analytics programmatically, without screen-scraping a dashboard.
Read the founder's full story →Architecture
Metrone runs on Cloudflare's global edge network. Events are ingested, validated, privacy-processed (IP hashing with rotating daily salt, geo resolution from network data), and written to PostgreSQL — all within milliseconds.
The tracking script (m.js) is ~2 KB, served from your domain via CNAME, and loads without cookies. SPA navigation is tracked automatically via pushState / replaceState interception.
Timezone-aware analytics use a pre-computed local_timestamp column — set by a database trigger at insert time — so read queries never need expensive timezone conversions.
Get in touch
Questions, partnerships, or feedback — we'd love to hear from you.
Contact us →