AnalyticsCAN I VIBECODE
FullStory?
BUILD FIRST
- A non-sensitive or read-only version of the core experience
- A manual review step wherever correctness or trust matters
- A narrow companion tool that removes a real piece of work
LEAVE FOR LATER
- Regulated decisions, payments, or irreversible actions
- Claims of correctness, compliance, or uptime that the MVP cannot support
- Sensitive integrations until security, support, and auditability are ready
RESEARCH NOTES
What this tool actually does.
FullStory is a behavioral data platform that combines session replay, product analytics, and debugging. A single JavaScript snippet captures every click, swipe, scroll, DOM change, and JavaScript console event automatically, with no manual event tagging. Its data engine maps digital properties to enable tagless analytics: dashboards, conversion funnels, retention charts, and journey mapping. Clicking any metric can open the corresponding session replays to see the 'why' behind the 'what'. Replays are high-fidelity, pixel-perfect recreations of the page (not video) and are fully inspectable, including DOM and console. The platform includes AI-generated one-click session summaries, privacy controls to mask sensitive content, and tools for product, engineering, and support teams. It supports mobile and desktop experiences and integrates with leading SaaS tools. The official docs describe it
RESEARCHED BUILD PROMPT
Give this to your coding agent.
Build a focused replacement prototype inspired by the FullStory workflow: capture real user sessions on a web app, make them searchable, and replay them with associated analytics. Implement a lightweight JavaScript SDK (a single small script) that automatically records clicks, scrolls, pointer moves, input events, DOM mutations, viewport size, and console/error log entries, without requiring manual event tagging. Transmit batches of compressed captured events over WebSocket or POST to a backend. Store events in a time-series/clickhouse-style store (use ClickHouse or Postgres+JSONB for the prototype). Build a session list UI with search by URL, user ID, date range, and error presence. Build an analytics dashboard generated from the captured event stream: page views, click heatmaps (aggregated by element selector), conversion funnel builder (define steps via URL/selector), retention chart by returning users, and a simple journey/flow map. Build a replay player that reconstructs the DOM from captured snapshot/mutation data using rrweb (or similar open-source library) with a timeline, console output panel, and no video encoding. Add privacy controls: configurable masking of element selectors and input values by default. Use TypeScript for the SDK, Node.js backend, and React frontend. Do not copy FullStory's brand, name, exact visual design, logo, or copy. Make the UI original and simple. Out of scope: mobile native SDKs, AI-generated session summaries, third-party integrations, multi-tenant SaaS features, and production-scale infrastructure. Write a README explaining setup, how the recording works, key limitations (e.g., iframe blocking, shadow DOM coverage, replay fidelity constraints, storage cost for long-term data).