H tspot Docs
Hotspot › Docs › Getting Started › Overview

Overview & Architecture

Hotspot is self-hosted product telemetry, session replay, and behavioral intelligence. See where users get stuck without sending data to third parties.

Hotspot is a lightweight, self-hosted product telemetry and behavioral analytics engine. With a single ~12KB script tag, it records full session replays, rage clicks, dead clicks, client errors, conversion funnels, and CRM customer journeys, keeping all data 100% on your own server.
💡
Privacy-First & Zero-PII: Unlike conventional analytics tools, Hotspot automatically redacts sensitive data, hashes user identifiers with tenant-scoped cryptographic salts, and does not require third-party cookie consent banners.

Why Hotspot? Comparison with Other Platforms

Modern engineering and product teams need deep behavioral visibility without vendor lock-in, data privacy violations, or sluggish page loads.

Capability Hotspot (Self-Hosted) Microsoft Clarity Hotjar / FullStory
Data Hosting 100% On-Prem / Cloud Microsoft Cloud Vendor Cloud
Script Size ~12 KB gzipped ~45 KB ~65+ KB
Zero-PII Hashing Deterministic HMAC-SHA256 Basic Masking Manual Redaction
Drop-Off Diagnostics Automated Root-Cause Manual Inspection Manual Inspection
VIP Customer Radar Real-Time Friction Alerts No Enterprise Addon
AI / MCP Integration Native MCP Server No No

System Architecture

Hotspot is designed with a lightweight, decoupled architecture that scales effortlessly on modest hardware:

1

Edge Telemetry SDK (hotspot.js)

Runs in browser background. Captures DOM mutations via rrweb, pageviews, clicks, dead clicks, rage clicks, JavaScript exceptions, and web vitals without impacting main thread performance.

2

Ingestion & Mutator Pipeline

Receives batched beacon payloads at /hotspot/batch/ and recording chunks at /hotspot/replay/. Computes live session state, user hashes, and funnel step progression on the fly.

3

In-Memory & Redis Persistence

Maintains sessions, site definitions, and events in Redis with automatic TTL pruning, multi-tenant site partitioning, and high-throughput write performance.

4

Dashboard, Analytics & MCP API

Zero-build standalone dashboard, customer journey drawer, funnel diagnostics, and streaming Model Context Protocol (MCP) server for AI assistants.