ClawVault

An Elephant Never Forgets

An open-source memory vault for AI agents. Store decisions, recall context, search by meaning. Never explain your project from scratch again.

Install with npm View on GitHub

terminal
$npm install -g @versatly/clawvault

“Let me explain the project again...”

Every time your AI agent restarts, it forgets everything. Decisions you made. Context you built. Lessons you learned together. You start from zero, every single time.

ClawVault changes that. Give your agent a memory vault where knowledge persists, accumulates, and grows.

How It Works

Simple CLI commands. Instant integration.

Store Memories

Your agent captures decisions, preferences, and learnings as structured memories. Each one categorized for fast retrieval later.

$clawvault store --category decisions \
--title "API Choice" \
--content "Chose REST over GraphQL for simplicity"

Search by Meaning

Ask natural language questions. Semantic search finds relevant memories even when exact keywords don't match.

$clawvault vsearch "what database did we pick?"

III

Checkpoint & Recover

Save working state before risky operations. If context dies, wake up with full awareness of where you left off.

# Before risky work
$clawvault checkpoint --working-on "refactoring auth"
# After restart
$clawvault wake

Six Commands. Infinite Memory.

Everything your agent needs to remember—stored, searchable, persistent.

Persistent Memory

Memories survive context resets, session ends, and restarts. Your agent remembers yesterday, last week, and last month.

Semantic Search

Find memories by meaning, not just keywords. Vector embeddings enable natural language queries across your vault.

Structured Categories

Organize into decisions, preferences, people, projects, and lessons. Everything has a place, retrieval is fast.

Context Recovery

Checkpoint before risky operations. If context dies mid-conversation, recover gracefully with full state.

CLI First

Designed for agent workflows. Simple commands that integrate seamlessly into any AI system or pipeline.

Open Source

MIT licensed, fully transparent. Human-readable markdown storage means you always own your data.

Common Questions

01What makes ClawVault different from other memory solutions?

ClawVault is designed specifically for AI agents, not humans. It stores memories in human-readable markdown files, supports both keyword and semantic search, and integrates seamlessly into any CLI-based workflow. Your data stays local and transparent.

02How does semantic search work?

ClawVault uses vector embeddings to understand the meaning behind your queries. Ask 'what database did we choose?' and it finds the relevant memory even if you never used those exact words when storing it.

03Will my data be used to train AI models?

No. ClawVault stores everything locally in markdown files. Your memories never leave your machine unless you explicitly share them. You have complete ownership and control.

04Does it work with any AI agent framework?

Yes. ClawVault is framework-agnostic. If your agent can run shell commands, it can use ClawVault. Works with OpenClaw, LangChain, AutoGPT, custom agents, or any tool that supports CLI integration.

05How do I recover from context death?

Run 'clawvault checkpoint' before risky operations. If your session dies, 'clawvault wake' restores your full working context. The handoff system ensures nothing is ever truly lost.

“The best agents don't just respond. They remember.”

Get Started