TencentCloud/TencentDB-Agent-Memory
TencentDB Agent Memory is a Hermes Agent plugin providing a 4-tier progressive and local-first memory pipeline to enhance AI agents by reducing token usage and improving task success rates.
Awesome Hermes Plugins › Memory Providers
MaxFreedomPollard/Compartment
Compartment is a durable memory layer for AI agents that runs entirely on the local machine. It ships a Hermes Agent plugin directory alongside an MCP server, so Hermes, Claude Code, Claude Desktop, OpenClaw and any other MCP-capable client read and write the same vault. Installation is a single pip command followed by compartment init, and compartment integrate hermes wires the agent and drops a /compartmentalize skill into the agent skills directory; a macOS .pkg bundles Python, the embedding model and every dependency for people who prefer not to use a terminal. Every byte at rest is AEAD-encrypted, including the embedding vectors, which the project notes are otherwise invertible back toward text. Deletion is cryptographic: destroying a record key removes the record irrecoverably. History is hash-chained so tampering is detectable, and the vault relocks on restart or power loss. A runtime guard aborts on any network attempt, and the project runs that check in CI on three operating systems. The embedding model ships inside the package, so search never leaves the machine; the authors report roughly 0.68 ms vector search and about 12 ms for the full hybrid pipeline. The vault is seeded at install with 6,718 curated facts covering physical constants and unit conversions, hardware specifications, operating system release names, network ports, file signatures, character encodings, shell and git internals, regular expressions, SQL, hashing and ISO country, currency and time codes; these are ordinary editable memories and one switch keeps them out of search results. Capture is ranked rather than filtered: decisions outrank preferences, preferences outrank machine details, and small talk is kept but ranked last. A hook records facts even when the model does not call the tool, and explicit relations with validity windows turn the store into a graph rather than a flat pile. Status, unlock, lock and recent memories are managed from a menu bar, notification area or window application.
https://github.com/MaxFreedomPollard/Compartment
Also in Memory Providers
MemoriLabs/Memori
Memori is an agent-native memory infrastructure, offering an LLM-agnostic layer that converts agent execution and conversation into structured, persistent state for production systems, including a ...
basicmachines-co/basic-memory
Basic Memory is a local-first, knowledge graph-based memory system for AI agents, providing persistent, bidirectional knowledge access via Markdown files and the MCP protocol, including a Hermes pl...
agentscope-ai/ReMe
A local-first Markdown memory service with a Hermes Agent plugin that turns conversations and documents into linked, searchable notes using wikilinks, BM25 and optional embedding retrieval.
mnemosyne-oss/mnemosyne
Mnemosyne is a zero-dependency, sub-millisecond AI memory system for Hermes Agents and other LLM frameworks, focusing on efficient, local, and private knowledge retention.
ClaudioDrews/memory-os
Memory OS is a 7-layer memory operating system for Hermes Agent, providing persistent memory, structured facts, fabric recall, an auto-curated wiki, and surgical context injection, running locally ...
zhangfengcdt/memoir
Memoir is a high-performance semantic memory system for AI agents, offering Git-like version control for AI memory management to combat context contamination, token costs, and memory drift.
itechmeat/open-second-brain
Open Second Brain provides an Obsidian-native, local-first memory layer for Hermes Agent, enabling deterministic learning and access across multiple AI runtimes.