LuizEduPP/rememb

v0.4.6 Bugfix

This release fixes issues for SREs watching stability and regressions.

Published 13d AI Agents & Assistants

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agent-memory ai-agents ai-memories ai-memory cherry-studio claude

+14 more

clawdbot context-persistence copilot cursor fastapi llm llm-memory mcp mcp-server memory model-context-protocol moltbot persistent-memory windsurf

ReleasePort's take

Light signal

editorial:auto 13d

Version 0.4.6 addresses cache overwriting issues, store initialization requirements, and UI formatting bugs. Metadata and documentation aligned across all targets.

Why it matters: Prevents semantic embedding cache overwriting on filtered operations and store state corruption during mutations. Operators using store edit/delete or web pagination should test in dev.

Summary

AI summary

Fixed embedding cache overwriting, store initialization checks, and entry‑formatting fallbacks in the web UI.

Changes in this release

Type	Severity	Summary	CVE
Dependency	Medium	Package and release metadata aligned to version 0.4.6 across all targets Package and release metadata aligned to version 0.4.6 across all targets Source: llm_adapter@2026-05-21 Confidence: low	—
Dependency	Medium	Package and release metadata aligned to version 0.4.6 across Python package, Docker image, and MCP registry. Package and release metadata aligned to version 0.4.6 across Python package, Docker image, and MCP registry. Source: granite4.1:30b@2026-05-21-audit Confidence: low	—
Bugfix
Bugfix	Medium	Store edit/delete operations require initialized root before mutating state Store edit/delete operations require initialized root before mutating state Source: llm_adapter@2026-05-21 Confidence: high	—
Bugfix	Medium	Semantic embedding cache no longer overwritten by filtered operations Semantic embedding cache no longer overwritten by filtered operations Source: llm_adapter@2026-05-21 Confidence: low	—
Bugfix	Medium	Entry formatting correctly distinguishes summary from full output Entry formatting correctly distinguishes summary from full output Source: llm_adapter@2026-05-21 Confidence: low	—
Bugfix	Medium	Web UI pagination safely handles invalid numeric input values Web UI pagination safely handles invalid numeric input values Source: llm_adapter@2026-05-21 Confidence: low	—
Bugfix	Medium	Semantic embedding cache persistence avoids being overwritten by filtered or section-scoped operations. Semantic embedding cache persistence avoids being overwritten by filtered or section-scoped operations. Source: granite4.1:30b@2026-05-21-audit Confidence: low	—
Refactor	Medium	README updated with current Web UI screenshots and information README updated with current Web UI screenshots and information Source: llm_adapter@2026-05-21 Confidence: low	—
Refactor	Low	README refreshed with current Web UI screenshots for memories, stats, and settings views. README refreshed with current Web UI screenshots for memories, stats, and settings views. Source: granite4.1:30b@2026-05-21-audit Confidence: low	—

Full changelog

Changed

Package and release metadata were aligned to version 0.4.6 across the Python package, Docker image install target, and MCP registry metadata.
README was refreshed to reflect the current Web UI and now includes real screenshots for the main memories view, stats view, and settings view.

Fixed

Semantic embedding cache persistence now avoids being overwritten by filtered or section-scoped operations, preventing stale subset embeddings from poisoning the shared on-disk cache.
Store edit and delete operations now require an initialized root before mutating state, avoiding partially initialized .rememb/ directories.
Entry formatting now distinguishes summary output from full output correctly, and the Web UI settings flow now falls back safely when numeric pagination inputs parse to invalid values.

View diff on GitHub

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

Share on X Share on Bluesky

Track LuizEduPP/rememb

Get notified when new releases ship.

About LuizEduPP/rememb

Persistent memory for AI agents with sectioned entries (project/user/context/etc), semantic search, CLI, and per-project scope. Local JSON, zero config, no server required.

All releases →