Local OpenTelemetry viewer for development. Run waggle next to your service, point any OTLP exporter at http://localhost:4318 (HTTP) or localhost:4317 (gRPC), and browse a Honeycomb-style trace waterfall, log explorer, metrics browser, and structured query builder in the same tab.

A single static Go binary — no CGO, no Docker required, no Node at runtime.

./waggle
# point your exporter at http://localhost:4318 (HTTP) or localhost:4317 (gRPC)

• Getting Started — install, first run, point an exporter at it.
• Configuration — every flag and environment variable.
• UI Guide — the explore page, query builder, trace waterfall, logs tail.
• Query Language — JSON AST the UI sends; build the same queries from your own tooling.
• MCP Server — explore your traces, metrics, and logs from Claude and other MCP clients.
• Tee Logs — mirror incoming log records to a terminal or file.
• Loadgen — synthetic OTel traffic for demos and benchmarks.

• Datasets and Schema — events vs metric_events, the wide-event model, meta.* virtual columns.
• OTLP Ingest — endpoints, content types, backpressure semantics.
• API Reference — the /api/* surface used by the UI.
• Performance — query timings on a 10M-row database.

• Limitations — what waggle does not do, and what to use instead.
• FAQ — short answers to common questions.
• Development — building from source, running tests, contributing.

Waggle is a local development tool. It runs on your laptop alongside the services you're working on and gives you a fast, signal-correlated view of what they emit. It is not a production observability backend.

If you want a hosted, multi-tenant platform with retention measured in months, sampling, RBAC, and alerting — use Honeycomb, Grafana Tempo/Loki/Mimir, or any other OTel-native vendor. Waggle complements those: it's the thing you run when you want to see a trace you just emitted immediately, in the same browser tab as everything else.

See Limitations for the full list.