Platform / Insight · preview

Athena

See what the runtime did. Open the number that looks wrong.

Console status: preview. Screens use synthetic data.

Athena · Insights
Athena · Insights7 days30 dayssample
Active runs
6
Completed · 7 days
1284
Error rate
1.8%
Review queue
4
Runs over time
completederrored
Outcomes
done86%gated9%error5%
When the work lands
RunKindStatusDurationStarted
invoice-intake #4112pipelinedone2m 14s12:41
invoice-intake #4111pipelinerunning—12:39
support-intakeagentdone48s12:31
invoice-intake #4110pipelineerror1m 02s12:18

What it is

Athena is the view over runs, review, and usage. A month of invoice intake should not be a spreadsheet someone exports on Friday. The dashboard is the same record the runtime wrote: completions, errors, what is waiting on a person, and when the work happened.

The console marks Athena as preview. The layout below is the dashboard — ranges, stat cards, a run chart, an outcome split, a weekday-by-hour heatmap, and a table of recent runs — on synthetic data. When you click through a rollup, the point is the row underneath, not a second tool.

What it is not

Athena does not decide a write and it does not define an agent. It shows what agentOS, pipelines, and agents already did. If a number matters, it should open into the run that produced it.

Who it is for

The operator who approves gates, and the lead who has to explain volume, failures, and review load without asking engineering for a query.

Where you get it

Platform. The console surface is in preview. It reads the org you already have. It is not a separate analytics subscription.

Pricing →

Monday after a bad Friday

Synthetic example. Not a customer.

Error rate on the 7-day card is up. You switch the range, see the spike in the afternoon, and open the runs table. One invoice pipeline failed on a webhook. The rest completed, including the ones that waited on review. You are looking at the runtime’s record, not a reconstructed report.

What changes when you have it

The review queue is visible

How many documents are waiting on a person, next to how many finished.

A spike has a row

From a range card to a single run, with status and duration.

Usage beside the work

Tokens, runs, documents, storage, and compute hours in the same place as the outcomes.

Inside Athena

  • Stat cards: active runs, completions, error rate, review queue
  • Runs-over-time and an outcome split
  • Weekday by hour heatmap
  • Recent runs with status
  • Usage meters for tokens, runs, documents, storage, and compute
  • Ranges: 24h, 7 days, 30 days, 90 days

How it works

1

Pick a range

24 hours, 7 days, 30 days, 90 days. The cards follow the range.

2

Read the shape

Runs over time, done versus gated versus error, and when in the week the work lands.

3

Open the row

A recent-runs table with live status. An anomaly is a specific run, not a mood.

4

Share the same view

Operators and viewers look at the org’s numbers, not a private export.

Athena. The charts are only useful if they are the runtime’s own record. Athena does not invent a second set of books.

The names next to it

See it on your workflow

Platform Start is the self-serve canvas. A first pipeline we stand up with you is delivery, priced apart from the subscription.