Introduction

Vision is The AI Business Observability Platform. We unify product usage, AI model inference costs, and financial unit economics into a single operational pane of glass.

Connect customer journeys, token spend, and subscription revenues to observe your real margins in real time.

Why Vision?

  • Unified Business Observability — Connect visitor behavior directly to LLM prompt executions and Stripe/Paddle payment events.
  • True Customer Unit Economics — Measure Gross Margin per User by automatically subtracting AI model compute cost from paid revenue.
  • Cost-to-Serve (CTS) Telemetry — Track how much each active account, prompt template, and AI feature costs to run across 40+ model providers.
  • Zero-Postgres High-Throughput Experiments — Run A/B tests and feature flags streaming directly through Kafka and Tinybird with sub-millisecond query latency and zero relational database load.
  • Native MCP (Model Context Protocol) Server — Ask AI questions about your live business health directly from Claude Desktop, Cursor, Antigravity, or custom agent frameworks.
  • Privacy-first & Lightweight — Under 3KB client tracker, no cookies, no third-party ad tracking, GDPR compliant out of the box.

Core Observability Pillars

  1. Visitor Tracking & Product Analytics — Pageviews, sessions, conversion funnels, user journeys, retention cohorts, and session replays.
  2. AI Cost Tracking — Full prompt traces, token volume, Time-To-First-Token (TTFT) latency, model benchmarks, and cache savings.
  3. Revenue Analytics — Subscription and transaction tracking via Stripe, Paddle, Polar, and Lemon Squeezy with user attribution.
  4. Unit Economics — Blended gross margins, Cost-to-Serve (CTS), customer payback velocity, and AI ROI ratios.
  5. Errors & Reliability — Catch prompt failures, rate limits, model timeouts, and client-side JavaScript errors.
  6. Ask AI via MCP — Connect your AI tools to query executive scorecards and live customer P&Ls.

Get Started

  1. Create a project in the Vision Dashboard to generate your tracking snippet and API keys.
  2. Embed the tracker in your frontend <head>.
  3. Route LLM calls through the Vision proxy or pass X-Vision-User-Id to link inference costs to user accounts.
  4. Connect your payment gateway to start observing customer gross margins.

Explore our Getting Started Guide or read about Unit Economics.