This release includes breaking changes for platform teams planning a safe upgrade.
Published 3mo
MCP Data & Storage
✓ No known CVEs patched
✓ No known CVEs patched in this version
Summary
AI summaryFixed re‑downloading of the embedding model on server restart and prevented crashes during initial download.
Full changelog
What's Changed
Bug Fixes
- Model caching: Fixed an
hf-hubissue that caused the 1.2 GB embedding model to be re-downloaded on every server restart. The cache directory is now respected correctly so the model is fetched only once. - SIGTERM / MCP timeout during download: MCP clients were timing out or the server was crashing with
SIGTERMwhile the embedding model was still being downloaded on first run. This is fixed by anensure_embedding_ready!wait loop that transparently blocks tool calls until the model is fully loaded, preventing both the timeout and the signal-induced crash. - Clippy warning: Fixed a minor Clippy lint (
&PathBufargument changed to&Path).
Maintenance
- Removed unused
HF_HOMEenvironment variable entries from dev configuration files. - Clarified Docker volume documentation — the correct cache path for the model inside the container is
/data/models.
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About pomazanbohdan/memory-mcp-1file
A self-contained Memory server with single-binary architecture (embedded DB & models, no dependencies). Provides persistent semantic and graph-based memory for AI agents.
Related context
Beta — feedback welcome: [email protected]