Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Addressed audit findings: added REDIS_PORT to metadata, clarified metadata-only indexing (no full-text), and ensured env.example.txt inclusion.
What This Skill Does
Mema provides a two-tier local memory system: a SQLite database that indexes document paths and tags, and a Redis store for ephemeral key-value state with a default 6-hour TTL. It lets agents track workspace files and pass context between sessions without relying on external services.
All data stays local, avoiding the privacy risks and network dependencies of cloud-based memory or context services.
When to use it
- Tagging and retrieving project documents by topic
- Passing session state between consecutive agent runs
- Building a local index of workspace knowledge paths
- Storing temporary values that expire after a work session
- Filtering indexed files by tag to surface relevant resources
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: 'Index my research notes and find all files tagged python'
- 1Runs mema init to create the SQLite schema if not already set up
- 2Runs mema index ~/notes/python-tips.md --tag python to record the file path and tag
- 3Runs mema list --tag python to query all indexed paths with the python tag
Returns a list of local file paths tagged 'python' from the SQLite index
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.