Agent-first observability.
Set up in one prompt.
Structured events, performance metrics, and conversion funnels — purpose-built to be driven by Claude Code, Codex, OpenClaw, Cursor, or whichever coding agent you use.
Get started
Add one config. Tell your agent. Done.
Add the MCP server to your editor, then your agent handles auth, project setup, and SDK integration.
claude mcp add --transport http --scope user owlmetry https://api.owlmetry.com/mcp \
--header "Authorization: Bearer YOUR_AGENT_KEY"Sign in above to get your key auto-filled
> Set up Owlmetry for this project and instrument the app with event tracking.
✓ Authenticated
✓ Project created
✓ App created
✓ SDK installed
✓ Instrumentation added
Done. Owlmetry is ready.Capabilities
Built for agents, usable by humans
Most observability tools are built for humans staring at dashboards. Owlmetry is built for agents making API calls.
Agent-Native API
Your coding agent sets up tracking, queries data, and diagnoses issues — no dashboard needed, though there's one if you want it.
Events
Know exactly what happened and when. Structured events with log levels, session tracking, and screen context give you a complete picture of every user journey.
Metrics
Find your slowest screens and flakiest network calls. Time any operation end-to-end and catch performance regressions before users notice.
Funnels
See where users drop off. Define multi-step funnels, compare conversion across app versions, and measure the impact of A/B experiments.
Issue Tracker
Error events cluster into issues automatically. Semver-aware regression detection, kanban workflow, and same-session burst aliasing so one bug doesn't triple-count.
Acquisition Attribution
See which Apple Search Ads campaigns drive your installs. The Swift SDK auto-captures attribution, and the dashboard ties campaign to user to revenue — no marketing SDK required.
User Feedback
Collect free-text feedback from your apps or your own frontend. Kanban board with status lifecycle, comments, and session-linked event replay for every submission.
Multi-Platform SDKs
Drop in a Swift or Node.js SDK and start collecting data in minutes. Batching, compression, and retry happen automatically.
Open Source
The entire codebase is open. Use our hosted service or run it on your own infrastructure — same product, your choice.
Use cases
Ask your agent. Get answers.
Your agent uses the Owlmetry CLI to query events, trace sessions, analyze funnels, and monitor performance — all from natural language.
>User 42 reported sync issues. What happened?
Querying errors for user 42...
Investigating surrounding events...
Tracing session timeline...
Found 5 sync errors in session. Background refresh failing with 409 conflict — stale local state after app update. Surrounding events show repeated retry loop.
>What app issues did we have overnight?
Pulling errors and warnings from the last 12 hours...
Grouping by message pattern...
17 errors overnight. 12 are "Payment gateway timeout" starting at 3:14 AM — Stripe had a 47-minute outage. 5 unrelated auth errors from a single device.
>Why is our onboarding conversion dropping?
Querying onboarding funnel for the last 7 days...
Comparing step drop-offs by app version...
Drop-off between "Create Account" and "Verify Email" jumped from 12% to 34%. Only affects v2.4.1 — email verification deeplink is broken on iOS 18.
>Is our photo upload getting slower?
Querying photo-upload metric for the last 7 days...
Comparing against the previous week...
p50 steady at 1.2s but p95 climbed from 3.8s to 6.1s this week. Failure rate up 2%. Large files (>5MB) are timing out on cellular connections.
>How is the checkout redesign experiment performing?
Querying checkout funnel segmented by experiment variant...
Comparing checkout completion time across variants...
Variant B converts 23% better than control (68% vs 55%) and checkout time dropped from 4.2s to 2.8s. Variant A shows no significant difference. Recommend shipping Variant B.
How it works
Your agent does the work
Add one config. Your agent handles everything else.
Add the MCP server
Paste a config snippet into your editor's MCP settings. One JSON block, one minute.
Agent sets you up
Your agent creates a project and app, configures API keys, and verifies the connection — all through MCP.
Agent instruments your code
The agent adds the SDK to your app and wires up event tracking, metrics, and funnels.
Pricing
Simple pricing. No surprises.
Start free. Scale when you're ready.
$0
forever
- 1 app
- 10,000 events per month
- Events, metrics & funnels
- Agent API keys + CLI
- Web dashboard
per month
- Unlimited apps
- Unlimited events
- Events, metrics & funnels
- Agent API keys + CLI
- Web dashboard
- Priority support
$0
forever
- Unlimited everything
- Your infrastructure, your data
- Single Postgres — no infra sprawl
- Full REST API
- Open source
Let your agent handle observability.
Your agent sets up the integration, monitors production, and acts on what it finds.