StagingDB checkingCommit: 5c5ec74eLast migration: checkingApp: staging.runtraice.comDB server: Cloud SQL via Unix socketLogical DB: traiceAuth: Staging OAuth or emailEmail: Brevo

Demo

Connect

Slack community

WhatsApp community

Product

People

Controls

ExperimentsPro+

GuardrailsTeam+

Workspace

Example workspace · Unified AI spend

Compare customer-facing AI costs with employee usage in one operating view. Actions are read-only.

Northstar Helpdesk AI

ProductSessions

Workspace

Run trace

Session
Run trace

Run trace

d72cd27e7b26f6654f5fd1b3de73e90f

Agent escalation-agent · Workflow wf_escalation_agent · Session conv_ef40781ed81e9668785d

Cost

$0.40685

136,538 tokens

Calls

4

No failed or retried calls

Duration

30.4 s

32% inside model calls, 41% in tools. The rest is code and waiting.

Timeline

Each bar starts when the call started and is as long as its latency. A call sits under the call, agent, or tool that made it when that one is stored; agent and tool rows come from OpenTelemetry spans and carry no cost. Open a row for its tokens, outcome, and ids.

Call0 ms7.61 s15.2 s22.8 s30.4 sLatency · cost

Started

Oct 27

Oct 27, 03:02:02 UTC