Workspace locked
INFERENCE ECONOMICS

Reveal your organization’s token usage.

Ork Meter reveals usage. Ork Meter Plus helps you get more value from the tokens you already pay for.

Reported API charges ↗
—
Selected feeds · Provider-reported USD
Token cost estimate ◈
—
Separate estimate · Never added to charges
Selected-feed tokens ≋
—
Input, output and disjoint cache categories
Connected sources ◎
—
Of 6 supported sources · Selected feeds

Selected-feed tokens by source

OBSERVED TOKENS

Source is the billing channel. Cursor usage stays under Cursor, even when its model is Claude.

What the numbers cover

EVIDENCE
Selected-feed events
—
Observed
Events with unknown cost
—
Needs evidence
Included-plan allocations
—
Separate
Clarity over false precision.
Connected feeds show observed activity. Amounts are observed subtotals, not complete bills. Unknown cost stays unknown; a zero requires supplied evidence. Subscription invoices and unconnected usage need separate evidence.

Usage by source and application

Filters apply to this table and its CSV download. Overview totals and native counter observations keep their existing scope.

Load a report to filter usage.

Source / modelApplicationCost centerTokensEstimateReported chargeFeed
Connect a reporting key to see your organization's usage.
ORK METER PLUS

Your tokens. More useful work.

CLASSIFICATION PILOT

Pre-classify a prompt with Ork Meter Plus using your preferences. Get a recommendation before choosing how to spend inference time and tokens.

Ork Meter · Planned $9.99/user/month or $100/year

Ork Meter Plus · Planned $20/user/month or $200/year total

Paid offers are not live. Sign in to see your account access. Marketplace access follows its separate purchase terms.

Connect an expiring Plus pilot reporting key to try classification. Savings require an applied change and outcome evidence.
YOUR PREFERENCES
TRY A PROMPT

A recommendation does not execute an inference request or modify your tools.

USAGE & OUTCOMES

Did the work pay off?

WORK-LEVEL ATTRIBUTION

Connect a work ID to consumption and commit evidence. Value uses your supplied baseline and labor assumptions.

Best available API cost
—
Reported first, then estimates
Net time recovered
—
Baseline minus AI and review effort
Capacity ROI scenario
—
Assumption-based · Excludes licenses/hosting
Net value / 1M tokens
—
All workflow tokens · Not cash savings
No implementation evidence loaded. Tokens alone do not establish useful work or code-token lineage.