# Traces

> Read every step a task took, from the console or the API. A trace is a task's step-by-step record. List them, read one, or send them to your own observability tools.

Canonical URL: https://harnessrouter.ai/docs/traces
Section: API reference

## Endpoints

`org` is required and must be your own organization. A workspace key lists its own workspace whatever it asks.

Trace endpoints:

```
GET /v1/traces?org={org}&limit=20&cursor=&harness=&workspace=&member=   tasks with traces, newest first
GET /v1/traces/{id}                                         one trace's summary
GET /v1/traces/{id}/events?chunk=0                          one chunk of its events
GET /v1/traces/{id}/all?compact=1                           every event; compact=0 keeps long strings whole
```

## Access

Tracing, in beta, is included with Scale and Enterprise, and is an add-on for the other plans. Without it, a trace shows its first five events in full and marks the rest as limited.

- [Pricing](https://harnessrouter.ai/pricing): The tracing add-on.

## Exports

- Send traces and logs to your own observability receivers from Organization, then Activity and exports. The organization's audit log is there too.
- Exports need the tracing add-on. Admins set them up, or people and keys an admin allows.
- Including prompts and outputs in an export needs an admin and a second confirmation.
- A workspace has its own exports only if an admin allows it.

- [Tracing and optimization](https://harnessrouter.ai/docs/tracing-and-optimization): Using traces to choose a harness and model.
