vizb

The root command (vizb) is the primary entry point. It parses benchmark output from multiple frameworks and generates interactive visualizations.

Usage

vizb [target] [flags]

Input Methods

File

vizb bench.txt -o output.html

Pass a file path as the first argument.

Stdin Pipe

go test -bench . | vizb -o output.html

Pipe benchmark output directly to vizb.

JSON Auto-Detect

# JSON bench events are auto-detected and converted for Go
go test -bench . -json | vizb -o output.html

Vizb automatically detects go test -bench -json format when using --parser go.

Multi-Language Parsers

Vizb supports Rust, JavaScript, and TypeScript benchmark frameworks in addition to Go. Use --parser to select the right one:

# Rust Criterion
cargo bench | vizb --parser rs:criterion -o output.html

# Rust Divan
cargo bench | vizb --parser rs:divan -o output.html

# Vitest
npx vitest bench | vizb --parser js:vitest -o output.html

# Tinybench
node bench.js | vizb --parser js:tinybench -o output.html

See the Parser Guide for full details on each parser’s metrics and configuration.

Flags

Flag	Short	Default	Description
--output	-o	(stdout)	Output file path. .json → JSON, else → HTML
--parser	-P	go	Benchmark parser: go, js:tinybench, js:vitest, rs:criterion, rs:divan
--name	-n	Benchmarks	Benchmark name
--description	-d	""	Benchmark description
--tag	-t	""	Tag identifier for release tracking
--group-pattern	-p	x	Pattern-based grouping (n/x/y with / or _)
--group-regex	-r	""	Regex-based grouping (named captures)
--scale	-S	linear	Scale type: linear or log
--sort	-s	""	Sort order: asc or desc (default: as-is)
--charts	-c	bar,line,pie	Chart types to generate
--show-labels	-l	false	Show value labels on charts
--filter	-f	""	Regex to include only matching benchmark names
--mem-unit	-M	B	Memory unit: b, B, KB, MB, GB
--time-unit	-T	ns	Time unit: ns, us, ms, s
--number-unit	-N	""	Number unit: K, M, B, T (default: as-is)

Examples

Basic usage

# Pipe benchmarks to HTML
go test -bench . | vizb -o output.html

# From a file
vizb bench.txt -o output.html

JSON output

# Generate JSON for later merging
vizb bench.txt -o data.json

# Then generate HTML from JSON
vizb data.json -o output.html

Grouping

# Pattern grouping (slash-separated 3D data)
vizb bench.txt -p n/x/y -o output.html

# Regex grouping
vizb bench.txt -r '(?<n>.*)/text=(?<x>.*)/level=(?<y>.*)' -o output.html

See the Grouping Guide for detailed examples.

Logarithmic scale

vizb bench.txt --scale log -o output.html

Filtering

# Include only benchmarks matching a pattern
vizb bench.txt -f "Sort" -o output.html

Tagging for release tracking

vizb bench.txt -o v1.json --tag v1.0 -n "MyBench"

See Merging Guide for tag-based comparison.