Open Source · MIT License

Persistent memory
for AI agents

Gnosys gives LLMs a knowledge layer that survives across sessions. Atomic markdown files, git-backed history, Obsidian-compatible. Works with any MCP client.

$ npm install -g gnosys-mcp copy

Everything agents need to remember

A complete memory system designed from the ground up for LLMs. No vector databases, no embeddings, no black boxes.

Atomic Markdown Files

Each memory is one markdown file with YAML frontmatter. Human-readable, git-diffable, and browsable in Obsidian.

Smart Discovery

FTS5-powered search with relevance keyword clouds. Agents describe what they're working on and get relevant memories back.

Layered Stores

Project, personal, and global knowledge layers. Project memory travels with the repo. Personal memory follows you everywhere.

Memory Lensing

Filter by category, tags, status, author, confidence, dates. Compound lenses combine criteria with AND/OR logic.

Git-Backed History

Every write auto-commits. View full version history, diff between versions, and non-destructively rollback any memory.

Obsidian Wikilinks

Cross-reference memories with [[wikilinks]]. Build a knowledge graph with backlinks, see connections, find orphaned links.

LLM Ingestion

Feed raw text and an LLM structures it into an atomic memory with title, tags, category, and relevance keywords.

Knowledge Timeline

Visualize how knowledge evolves over time. Group by day, week, month, or year. Track creation and modification patterns.

Bulk Import

Import thousands of records from CSV, JSON, or JSONL. LLM-powered ingestion generates keyword clouds automatically. Batch commits, dedup, and resume support.

Cross-Platform

Runs on macOS, Linux, and Windows. Pure Node.js with zero platform-specific dependencies. Install once, works everywhere.

Add to your editor in 2 minutes

Gnosys is an MCP server. Point your AI client at it, and your agent gains persistent memory.

claude_desktop_config.json
{
  "mcpServers": {
    "gnosys": {
      "command": "npx",
      "args": ["gnosys-mcp"],
      "env": {
        "ANTHROPIC_API_KEY": "sk-..."
      }
    }
  }
}
.cursor/mcp.json
{
  "mcpServers": {
    "gnosys": {
      "command": "npx",
      "args": ["gnosys-mcp"],
      "env": {
        "ANTHROPIC_API_KEY": "sk-..."
      }
    }
  }
}
terminal
$ claude mcp add gnosys npx gnosys-mcp
~/.codex/config.toml
[mcp.gnosys]
type = "local"
command = ["npx", "gnosys-mcp"]

[mcp.gnosys.env]
ANTHROPIC_API_KEY = "your-key-here"
opencode.json
{
  "mcp": {
    "gnosys": {
      "type": "local",
      "command": ["npx", "gnosys-mcp"],
      "env": {
        "ANTHROPIC_API_KEY": "your-key-here"
      }
    }
  }
}
.cursor/rules/gnosys.mdc
---
description: Gnosys persistent memory
alwaysApply: true
---

# Gnosys Memory System

## Retrieve memories
- At task start, call gnosys_discover with keywords
- Load results with gnosys_read
- Trigger on: "recall", "remember when", "what did we decide"

## Write memories
- Trigger on: "remember", "memorize", "save this", "don't forget"
- Also write on decisions, preferences, specs, post-task findings

## Key Tools
gnosys_discover → find memories  gnosys_add → write
gnosys_read     → load content   gnosys_update → modify
gnosys_hybrid_search → best search
gnosys_ask → Q&A with citations
… plus 25 more tools (maintain, history, graph, etc.)
CLAUDE.md
# Gnosys Memory

This project uses Gnosys for persistent memory via MCP.

## Read first
- At task start, call gnosys_discover with keywords
- Load results with gnosys_read
- On "recall", "remember when", "what did we decide" → search first

## Write automatically
- On "remember", "memorize", "save this" → call gnosys_add
- Decisions/preferences → commit to decisions/
- Specs → commit BEFORE starting work
- After implementation → commit findings

## Key tools
| Action | Tool |
| Find   | gnosys_discovergnosys_read |
| Search | gnosys_hybrid_search, gnosys_ask |
| Write  | gnosys_add, gnosys_add_structured |
| Update | gnosys_update, gnosys_reinforce |
1

Install the package

One command: npm install -g gnosys-mcp. Zero config needed to start.

2

Add to your MCP client

Works with Claude Desktop, Cursor, Claude Code, Codex, OpenCode, or any MCP-compatible client on macOS, Linux, or Windows. Just add the server config.

3

Initialize a store

Ask your agent to run gnosys_init in your project. It creates a .gnosys/ directory with git tracking.

4

Your agent remembers

Decisions, architecture choices, conventions, requirements — all persisted as atomic markdown files your agent can discover and reference.

No black boxes

Every memory is a plain markdown file you can read, edit, and browse in Obsidian. Git tracks every change. The filesystem is the source of truth.

AI Agent
Claude, Cursor, or any
MCP-compatible client
Gnosys MCP
17 tools for memory
operations over stdio
.gnosys/
Markdown + YAML
in a git repo
Obsidian
Browse, search, graph
your knowledge vault

Give your agents a memory

Open source, MIT licensed. Built for developers who want their AI to remember what matters.