Universal memory for any AI agent

Give your AI agent
a memory that persists

Memorius is a self-contained, agent-agnostic memory system with hierarchical knowledge organization, pluggable embeddings, and neuroscience-inspired consolidation. Works with 7 AI coding agents out of the box.

$ pip install memorius
v0.9.0 MIT licensed Python 3.11+ 7 agent integrations MCP + REST + Hooks
Capabilities

Built for the way agents actually work

Drop-in memory that any AI agent can use. No lock-in, no per-agent rewrites — one vault, many frontends.

Drop-in for any MCP client

Point any MCP-compatible client at memorius serve and instantly get memory, semantic search, and diary tools. No SDK to embed.

MCP Server

Hierarchical knowledge

Organize memory with a Vault → Shelf → Folder → Note hierarchy. Mine conversation transcripts into structured memories automatically.

Learn the hierarchy

Pluggable embeddings

ChromaDB default with bundled ONNX, sentence-transformers for higher quality, or OpenAI for production. Swap with one config change.

Configuration

7 agent integrations

Claude Code, Codex CLI, Gemini CLI, OpenCode, Pi, OpenClaude, OpenClaw — auto-detected from hook payloads, no config needed.

All agents

Neuroscience consolidation

Interference-based forgetting, engram maturation, and neuromodulator gating model how biological memory systems decay and consolidate. Typed graph edges with spreading activation.

Memory engine

REST API + Obsidian sync

FastAPI HTTP server for any non-MCP client. Bidirectional Obsidian vault sync with YAML frontmatter and folder hierarchy preservation.

REST API
Documentation

Everything you need to work with Memorius

Each concern lives on its own page. Start with getting started, then dive into the exact surface you're wiring up.

Getting Started

Install, initialize, store your first memory, and understand the Vault → Shelf → Folder → Note hierarchy.

Read

CLI Reference

Every subcommand — store, search, mine, extract, consolidate, prune, repair, doctor, and the full flags on each.

Browse commands

MCP Server

Point any MCP client at memorius serve. All 29 tools, JSON-RPC over stdio, client config snippets.

See the tools

Agent Hooks

Auto-detecting lifecycle hooks for 7 agents. 10 action types, the detection priority, and the memorius-hook CLI.

Wire up hooks

REST API

FastAPI server for non-MCP clients. Full endpoint table, API key auth, rate limits, daemon mode, TLS.

Reference

Architecture

How the vault engine, search pipeline, embeddings, vector backends, and knowledge graph fit together.

Understand it

Agent Compatibility

Which agents are supported, the hook event each emits, detection priority, and how to plug in your own.

Supported agents

Memory Engine

Temporal decay, consolidation, pruning, knowledge graph edges, fact-checking, and the repair / doctor health tools.

The engine

Agent Skill Install

One npx command installs the pre-built agent skill that teaches Hermes-compatible agents to use memorius proactively.

Install the skill

Ready to give your agent
a memory that persists?

Install in one command, wire up MCP or hooks, and start storing memories immediately. No cloud, no API keys, no lock-in.