Files

T

History

osobhandClaude Opus 5.5 d6e426e6d5 fix(agent): fail to open a store whose /meta has an unreadable attribute

Group::attrs now leaves out an attribute it cannot decode. The agent read
its settings through it, so a store whose float16 (or compression,
quantized_index, WAL mark, signature...) attribute could not be decoded
opened with the default in its place, and no error. /meta is now read
with attrs_with_errors and any unreadable attribute is a Schema error,
as it was before attrs became tolerant. Found by the adversarial review.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

2026-09-25 22:43:53 -05:00

benches

bench: re-run every stale BENCHMARKS.md section, dated and traced

2026-09-24 23:45:15 -05:00

fuzz

test(agent): WAL property tests, crash-recovery matrix, WAL fuzz target

2026-09-19 06:43:34 -07:00

src

fix(agent): fail to open a store whose /meta has an unreadable attribute

2026-09-25 22:43:53 -05:00

tests

fix(agent): fail to open a store whose /meta has an unreadable attribute

2026-09-25 22:43:53 -05:00

Cargo.toml

feat(agent): Ed25519-signed checkpoints

2026-09-25 10:13:34 -05:00

README.md

docs: fix stale package names, counts, and CLI subcommands

2026-06-03 11:17:04 +00:00

README.md

clawhdf5-agent

HDF5-backed persistent memory store for on-device AI agents.

Built on clawhdf5, clawhdf5-agent provides a vector-searchable memory backend optimized for edge AI workloads. Store embeddings, text chunks, and metadata in a single HDF5 file with SIMD-accelerated similarity search.

Features

Persistent vector store in HDF5 format
Cosine similarity and L2 distance search
SIMD-accelerated via clawhdf5-accel (AVX2, NEON)
Optional GPU acceleration via clawhdf5-gpu
Memory-mapped access for large stores
f16 storage support for compact embeddings

Usage

[dependencies]
clawhdf5-agent = "2.1.0"

License

MIT