YourMemory

YourMemory

+16pp better recall than Mem0 on LoCoMo. 100% stale memory precision. Biologically-inspired memory decay for AI agents.

Persistent memory for Claude that works like human memory — important things stick, forgotten things fade, outdated facts get demoted automatically.

Early stage — feedback and ideas welcome.

Benchmarks

Evaluated against Mem0 (free tier) on the public LoCoMo dataset (Snap Research) — 10 conversation pairs, 200 QA pairs total.

Full methodology and per-sample results in BENCHMARKS.md.Read the writeup: I built memory decay for AI agents using the Ebbinghaus forgetting curve

How it works

Ebbinghaus Forgetting Curve

base_λ      = DECAY_RATES[category]
effective_λ = base_λ × (1 - importance × 0.8)
strength    = importance × e^(-effective_λ × days) × (1 + recall_count × 0.2)
score       = cosine_similarity × strength

Decay rate varies by category — failure memories fade fast, strategies persist longer:

Importance additionally modulates the decay rate within each category. Memories recalled frequently gain recall_count boosts that counteract decay. Memories below strength 0.05 are pruned automatically.

Setup

Zero infrastructure required — uses SQLite out of the box. Two commands and you're done.

1. Install

pip install yourmemory

All dependencies are installed automatically. No clone, no separate download steps needed.

2. Wire into your AI client

The database is created automatically at ~/.yourmemory/memories.db on first use. No .env file needed.

Claude Code

Add to ~/.claude/settings.json:

{
  "mcpServers": {
    "yourmemory": {
      "command": "yourmemory"
    }
  }
}

Reload Claude Code (Cmd+Shift+P → Developer: Reload Window).

Cline (VS Code)

1. Open VS Code → Cline extension → click MCP Servers icon
2. Click "Edit MCP Settings"
3. Add the following (uses python3 directly — no PATH issues):

{
  "mcpServers": {
    "yourmemory": {
      "command": "python3",
      "args": ["-m", "memory_mcp"]
    }
  }
}

5. Save — Cline will detect the server automatically.

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "yourmemory": {
      "command": "yourmemory"
    }
  }
}

Restart Claude Desktop.

Any MCP-compatible client

YourMemory is a standard stdio MCP server. The command is simply yourmemory. Add it to any client that supports MCP servers using the same pattern above.

3. Add memory instructions to your project

Copy sample_CLAUDE.md into your project root as CLAUDE.md and replace:

• YOUR_NAME — your name (e.g. Alice)
• YOUR_USER_ID — used to namespace memories (e.g. alice)

Claude will now follow the recall → store → update workflow automatically on every task.

PostgreSQL (optional — for teams or large datasets)

If you have PostgreSQL + pgvector, create a .env file:

DATABASE_URL=postgresql://YOUR_USER@localhost:5432/yourmemory

The backend is selected automatically — postgresql:// in DATABASE_URL → Postgres + pgvector, anything else → SQLite.

macOS

brew install postgresql@16 pgvector && brew services start postgresql@16
createdb yourmemory

Ubuntu / Debian

sudo apt install postgresql postgresql-contrib postgresql-16-pgvector
createdb yourmemory

One-liner setup script (macOS/Linux): bash scripts/setup_db.sh handles install + DB creation automatically.

MCP Tools

store_memory accepts an optional category parameter to control decay rate:

# Failure — decays in ~11 days (environment changes fast)
store_memory(
    content="OAuth for client X fails — redirect URI must be app.example.com",
    importance=0.6,
    category="failure"
)

# Strategy — decays in ~38 days (successful patterns stay relevant)
store_memory(
    content="Cursor pagination fixed the 30s timeout on large user queries",
    importance=0.7,
    category="strategy"
)

Example session

User: "I prefer tabs over spaces in all my Python projects"

Claude:
  → recall_memory("tabs spaces Python preferences")   # nothing found
  → store_memory("Sachit prefers tabs over spaces in Python", importance=0.9, category="fact")

Next session:
  → recall_memory("Python formatting")
  ← {"content": "Sachit prefers tabs over spaces in Python", "strength": 0.87}
  → Claude now knows without being told again

Decay Job

Runs automatically every 24 hours on startup — no cron needed. Memories below strength 0.05 are pruned.

REST API

# Store
curl -X POST http://localhost:8000/memories \
  -H "Content-Type: application/json" \
  -d '{"userId":"u1","content":"Prefers dark mode","importance":0.8}'

# Retrieve
curl -X POST http://localhost:8000/retrieve \
  -H "Content-Type: application/json" \
  -d '{"userId":"u1","query":"UI preferences"}'

# List all
curl "http://localhost:8000/memories?userId=u1"

# Update
curl -X PUT http://localhost:8000/memories/42 \
  -H "Content-Type: application/json" \
  -d '{"content":"Prefers dark mode in all apps","importance":0.85}'

# Delete
curl -X DELETE http://localhost:8000/memories/42

Stack

• PostgreSQL + pgvector — vector similarity search
• sentence-transformers — local embeddings (all-mpnet-base-v2, 768 dims, no external service needed)
• FastAPI — REST server
• APScheduler — automatic 24h decay job
• MCP — Claude integration via Model Context Protocol

Architecture

Claude Code
    │
    ├── recall_memory(query)
    │       └── embed → cosine similarity → score = sim × strength → top-k
    │
    ├── store_memory(content, importance, category?)
    │       └── is_question? → reject
    │           category: fact | assumption | failure | strategy
    │           embed() → INSERT memories
    │
    └── update_memory(id, new_content)
            └── embed(new_content) → UPDATE memories

PostgreSQL (pgvector)
    └── memories
        ├── embedding vector(768)
        ├── importance float
        ├── recall_count int
        └── last_accessed_at

Dataset Reference

Benchmarks use the LoCoMo dataset by Snap Research — a public long-context memory benchmark for multi-session dialogue.

Maharana et al. (2024). LoCoMo: Long Context Multimodal Benchmark for Dialogue. Snap Research.

License

Copyright 2026 Sachit Misra

Licensed under the Apache License, Version 2.0.You may use, modify, and distribute this software freely with attribution.Patent protection included — contributors cannot sue users over patent claims.