HDF5 Core Implementation
A complete HDF5 binary format reader built from the
spec,
in 63 KB(min) of JavaScript. No native libraries. No WebAssembly. Just bytes.
Supports superblock v0–v3, object headers v1/v2, layout messages v1–v4,
dataspace v1/v2, filter pipeline v1/v2, and attribute messages v1–v3.
Browser-based — files under 200 MB recommended.
Try it: Drop any .h5 file,
browse the tree, click a dataset to read its data.
superblock v0–v3
object headers v1/v2
symbol tables
link messages
contiguous data
chunked data
deflate / zlib
b-tree v1
fixed array index
float32/64 int8–32
gate pipeline viz
structural search
value search + elimination
byte inspector
attribute display
string datasets
heatmap / sparkline
sub-op tracing
extensible array idx
B-tree v2 idx
shuffle filter
fletcher32
compact layout
big-endian swap
5-gate pipeline
header hex inspector
vlen strings
global heap
compound types
implicit chunk idx
Runs entirely in your browser. No data leaves your machine.