LaunchKit · 2026
Back to Skills

braindb

Persistent, semantic memory for AI agents.

0
582 downloads
by @chair4ce

Setup & Installation

openclaw skills install @chair4ce/braindb

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install braindb

Version History

v0.5.2Feb 15, 2026 - Version note by skill developer:

Fix display name

What This Skill Does

BrainDB gives AI agents persistent memory that survives session resets and context compaction. It automatically extracts facts from conversations and retrieves relevant ones on demand, with no manual commands required. Memory persists across restarts with 98% recall accuracy and 12-20ms average query latency.

768-dimensional semantic search finds conceptually related memories rather than relying on keyword matching or manual file lookups.

When to use it

  • Remembering project architecture decisions across multi-day coding sessions
  • Recalling user preferences without re-explaining them each session
  • Importing existing MEMORY.md files and daily notes into semantic search
  • Persisting troubleshooting history and past resolutions
  • Maintaining long-running project context that outlasts the AI context window

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User says: 'We switched from MongoDB to PostgreSQL last week because of the relational data requirements.'

AGENT
  1. 1OpenClaw detects a notable fact and forwards it to BrainDB
  2. 2BrainDB encodes the fact as a 768-dim embedding and stores it in Neo4j
  3. 3In a new session, user asks about the database setup
  4. 4BrainDB retrieves the stored decision via semantic similarity search
  5. 5AI receives the prior context and includes it in its response
OUTPUT

AI correctly references the PostgreSQL decision and its rationale without being re-told

Requirements

Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.

Docker (runs the Neo4j, embedder, and gateway containers; requires ~4 GB RAM and ~3 GB disk)