First critical checks
Evaluating walkerOS? These five checks each take a few minutes, run on your own machine, and need no account. Together they test the claims everything else builds on: your tracking is code you can see, prove, and take with you.
1. See a live event, end to end
Send one event and watch it arrive. The quickstart
prints your first event from Node in under five minutes. Or run the
Docker quickstart and POST an event to
/collect: it appears in the runner log of a container on your machine.
2. Read your tracking as code
A bundled pipeline is one flow.json: sources, transformers, destinations, and
mapping as plain JSON you can diff in a pull request and roll back in git. Open
any example and read the whole
setup in one block. No UI state, nothing hidden.
3. Let an agent build a flow
The walkerOS MCP runs locally without an account. Ask your
AI assistant for a flow; it generates the flow.json, validates it against the
schema, and simulates an event through it before you ever deploy. The output is
a machine-checked artifact, not a demo.
4. Break it on purpose
Define a contract, give a step example
an event that violates it, and watch walkeros validate --strict fail with the
exact path and message, at design time, before anything ships. See
validate.
Compare that with a tag manager silently dropping fields in production.
5. Prove you can leave
Bundle the same flow.json and run it on your own infrastructure, in Docker or
directly in Node. Deploy shows each path.
Your events land in destinations you configured, so there is nothing to export
and nothing to migrate away from.
Next steps
- Quickstart: send your first event now
- GA4 ecommerce: a real destination with mapping, verified in DebugView