Team workspace

Switch orgs and gateways, filter the Sessions list, and read the aggregate dashboard for your team.

A paper console workspace is your team’s view of captured agent activity — every session anyone on the team ran, in one place.

Switch orgs and gateways

If you belong to more than one organization or have access to multiple gateways, switch between them in the console header. The Sessions view refreshes to show runs captured under the selected org and gateway.

Sessions dashboard

The Sessions dashboard gives you an aggregate view of all agent activity across your team. Use it to understand how AI work is being consumed at scale, not just session by session.

Filter the Sessions list

Three controls narrow what you see:

  • Scope — toggle between every session on your team and just your own.
  • Time range — focus on the window of activity you care about.
  • Search and filter — find a specific run, surface failed sessions, or sort by cost.

Sessions

Total Spend

$88.70

$11.09 avg

Sessions

8

1 running now

Tokens Used

67.2M

8.4M avg

Success Rate

71%

5/7 complete

a3f8c21dRunning
wire outcome detector into cursor cache$18.40
brian(You)
Claude Fable 5+2
d91b04eeCompleted
triage flaky gateway e2e$2.10
maya
Claude Haiku 4.5
7c25e9baCompleted
migrate span search to direct reads$31.75
brian(You)
Claude Fable 5+1
f04a77c3Completed
scaffold billing exporter cron$6.88
devon
GPT-5.2 Codex+1
b8e13a56Failed
refactor gateway auth middleware$4.12
maya
Claude Sonnet 5
c47d90f1Completed
bump nixpkgs and rebuild stereOS image$9.30
devon
Claude Opus 4.8+1
e62a4d8cFailed
add retry budget to overdub uploader$3.55
brian(You)
Claude Fable 5
91b7f3e0Completed
prototype skills marketplace page$12.60
maya
Claude Sonnet 5+1
Live demo — switch scope, pick a window, or search — the list and stats follow

Each row shows the session’s status, label, model, author, project, cost, tokens, duration, and when it started.

See which model ran each session

The Model column, next to the session name, shows the model that did most of the work in each session — brand logo and name — so you can scan a page of sessions without opening one. When more than one model served a session, the chip labels the cost-dominant model and adds a +n badge for the rest; hover it for the full breakdown of what each model cost. The per-model costs are exact, not estimates.

a3f8c21dRunning
wire outcome detector into cursor cache$18.40
brian(You)
Claude Fable 5+2
d91b04eeCompleted
fix flaky export-sessions test$2.10
maya
Claude Sonnet 5
7c25e9baCompleted
migrate span search to direct reads$31.75
brian(You)
Claude Opus 4.8+1
f04a77c3Completed
scaffold billing exporter cron$6.88
devon
GPT-5.2 Codex+1
58e1d20fFailed
triage grove sync failure$1.04
maya
Claude Haiku 4.5
Live demo — hover a model chip with a +n badge for the cost breakdown

The same chip sits in the session detail header with the same hover breakdown.

Export recent sessions

Session export is available on the Scale plan. On it, the Sessions list header has an Export last 30 days control that bundles every session from the past 30 days into a single JSONL file (sessions-last-30-days-<date>.jsonl). To export one session, use the Export button on its detail page.

Sessions

Total Spend

$630.83

$4.60 avg

Tokens Used

479.53M

477.96M in / 1.57M out

Agent Time

20h 15m

8m avg

Tool Calls

2337

17 avg

Success Rate

96%

131/137 complete

a3f8c21dRunning
wire outcome detector into cursor cache$18.40
brian(You)
d91b04eeCompleted
fix flaky export-sessions test$2.10
maya
7c25e9baCompleted
migrate span search to direct reads$31.75
brian(You)
f04a77c3Completed
scaffold billing exporter cron$6.88
devon
Live demo — click the download button in the header to export the last 30 days

Open a session

Click any row to open the session detail page for that run. From there you can read the full run, export it, or delete it.

Copied to clipboard