Free, open source

Four tools.One binary.

Analytics, APM, error tracking, and session replay — in a single self-hosted binary. One tool instead of four. Your data on your server. No per-seat pricing, no data caps.

brew install useteploy/tap/observe
Read the docs →

What you're paying today

5M events/mo, 10 hosts, 50k errors/mo

PostHogProduct analytics (4M events past free tier at $0.00005/event)
~$200/mo
SentryError tracking (Business plan, 50k errors included)
$80/mo
DatadogAPM (10 hosts at $31/host/mo)
$310/mo
UmamiWeb analytics (Pro cloud plan)
$20/mo
4 tools, 4 dashboards, data doesn't connect
~$610/mo

What you could have instead

teploy Observe
Analytics + APM + Errors + Sessions
$0
self-hosted, any scale
1 dashboard for everything
Your server, your data
Click error -- see trace -- see user journey
No data limits, no per-seat pricing
1 dashboard, 1 install, 0 bills
Save $7,300–$45,000+/yr depending on scale

Everything observability, nothing else

Each vertical replaces a dedicated tool. They share a unified data layer, so insights connect automatically.

Replaces PostHog + Umami

Web Analytics

Privacy-first, cookie-free analytics. No consent banners needed.

  • Pageviews, sessions, unique visitors
  • Custom events with properties
  • Funnels and conversion tracking
  • UTM campaign tracking
  • Geographic and device breakdown
  • Real-time active visitors
  • Referrer and channel analysis
Replaces Datadog + SigNoz

APM & Tracing

OpenTelemetry-native distributed tracing with automatic service maps.

  • Distributed traces (OTLP compatible)
  • Service dependency maps
  • Latency percentiles (p50/p95/p99)
  • Error rates per endpoint
  • Slow query detection
  • Request/response inspection
  • Infrastructure metrics
Replaces Sentry

Error Tracking

Crash reports with full context. Source maps, stack traces, breadcrumbs.

  • Automatic error grouping
  • Stack traces with source maps
  • Release tracking
  • Error frequency and trends
  • Issue management (open/resolved/ignored)
  • User impact analysis
  • Alert on new or regressed issues
The whole point

Unified Dashboard

The reason to use one tool instead of four. Everything is connected.

  • Click an error -- see the trace that caused it
  • Click a slow page -- see the API calls behind it
  • Click a user -- see their journey, errors, and sessions
  • One query language across all data
  • Custom dashboards combining any metric
  • Saved views and team sharing
  • Time-synced investigation across all verticals

One binary. Seriously.

Observe is a single Go binary backed by Nucleus, our multi-model database. Download, run, point your app at it.

1

Install

curl -sL https://raw.githubusercontent.com/useteploy/teploy-observe/main/scripts/install.sh | sh

Single binary. No Docker, no Kubernetes, no config files.

2

Add the tracker

<script src="observe.js" data-site="YOUR_ID"></script>

2KB script. Cookie-free. No consent banner needed.

3

See your data

open http://your-server:8080

Dashboard is live. Real-time visitors, analytics, errors, traces.

How Observe compares

ObservePostHogSentryDatadogUmami
Web analyticsYesYesNoLimitedYes
APM / TracingYesNoLimitedYesNo
Error trackingYesNoYesLimitedNo
Custom eventsYesYesYesYesYes
Self-hostedYesYesYesNoYes
Single binaryYesNo (Docker)No (Docker)No (SaaS)No (Docker)
Cookie-freeYesNoN/AN/AYes
OTLP compatibleYesNoNoYesNo
Unified dashboardYesAnalytics onlyErrors onlyInfra focusAnalytics only
Price (5M events/mo, 10 hosts)$0~$200/mo$80/mo$310/mo$20/mo
Verified pricing at 5M events/mo, 10 hosts. PostHog free tier covers 1M events — usage-based after that. Datadog at $31/host/mo APM. Costs compound significantly at scale.

Run it yourself

One binary. Your server. No accounts, no event caps, no per-seat pricing.