TokenMeter

A one-line context-window meter: used/limit token counts, a utilization bar, and a percentage (Context: 78k/200k [####......] 39%). From warningThreshold of the limit (80% by default) it adds NEAR LIMIT, and from errorThreshold (95% by default) OVER LIMIT — words, not color alone, so the state reads in monochrome.

Import: package:fleury/fleury.dart, or package:fleury/fleury_core.dart in browser code.

Usage

TokenMeter(
  usage: const TokenUsage(contextUsed: 128000, contextLimit: 200000),
  label: 'Context',
  width: 20,
)

Constructors

TokenMeter()

Parameter	Type	Default	Description
usage:	TokenUsage	required	Token and context-window metrics to display.
label:	String	'Context'	Visible and semantic label placed before the usage value.
width:	int	10	Number of cells used by the utilization bar; zero hides the bar.
warningThreshold:	double	0.8	Context-usage ratio at which the near-limit state begins.
errorThreshold:	double	0.95	Context-usage ratio at which the over-limit state begins.

Source

TokenMeter is defined in packages/fleury/lib/src/catalog/model_status_bar.dart.

Category: Agent surfaces · All widgets

Live demo

tokenmeter_example.dart

Restart Run Expand

Example · editable Revert Format

Widget example() => const TokenMeter(
  usage: TokenUsage(
    input: 8200,
    output: 1400,
    contextUsed: 128000,
    contextLimit: 200000,
  ),
  label: 'Context',
  width: 20,
);

Live preview

Live

ⓘ how this demo runs ⛶ Expand