Community metrics

Open project.Transparent numbers.

Explore downloads, GitHub stars and reported token savings, with their sources and calculation methods.

Public sourcesReported totalsExplicit assumptions

The reported total

Reported savings.

Token savings are reported voluntarily. Cost and energy figures are calculated estimates, not verified bills or electricity measurements.

Reported tokens saved
184.1B
Estimated cost avoided
$460,684
Estimated energy saved
20.5 MWh

Snapshot: . An unavailable source keeps the dated snapshot.

The community ledger

Every contribution. In the open.

Explore the public reports behind the totals. Compare reported token reductions, compression and estimated impact, then open a contribution to inspect its details.

3,233public reports
Public community reports. Self-reported data; cost and energy are estimates.
RankContributorTokens savedCompressionEst. energyEst. cost avoidedLast update
1frpboy20B
70.7%
2.2 MWh$49,998
2Anonymous9.5B
62.6%
1.1 MWh$23,738
3Anonymous9.4B
92.1%
1.0 MWh$23,525
4tmk5.5B
87.1%
605.8 kWh$13,631
5Anonymous3.5B
85.1%
393.5 kWh$8,853
6Anonymous3.4B
86%
375.5 kWh$8,450
7Anonymous2.5B
43.3%
278.5 kWh$6,267
8Anonymous2.4B
40.9%
265.8 kWh$5,981
9Anonymous2.2B
53.2%
243.1 kWh$5,470
10Anonymous2.2B
63.3%
239.7 kWh$5,394

Cost and energy are derived estimates. Service-flagged reports stay marked as unverified; a rank is not an audit or a measure of task quality.

The methodology

How the numbers are calculated.

The adoption snapshot records 324,252 download events and 3,946 GitHub stars. Neither is a count of active customers.

Downloads & repository interest
Downloads aggregate crates.io downloads, four npm distribution counters and GitHub release binary downloads. Checksum and signature files and self-update runs are excluded; repeat installs and short-lived environments still count. Stars are the repository’s public star count.
Community savings
The global leaderboard totals cover voluntary reports, including entries beyond the visible page. Reports are not independent verification. Estimated costs depend on the model prices used by each ledger.
The energy assumption
Reported tokens saved × 0.4 joules ÷ 3,600 = watt-hours. This fixed conversion is an illustrative model, not measured electricity. Real energy depends on model, hardware and the type of inference work avoided.

Your own evidence

Measure your own workflow.

Community reports show what others observed. Use your local measurements and a comparable task to understand the impact in your own work.

Your local report
lean-ctx gain --json

Inspect recorded context usage on your device. This command does not publish a report or add you to the community table.

Start with LeanCTX

Put context to work.

Set up LeanCTX for your tool and evaluate the difference on a real task.

Local-first. Model-agnostic. Your context.