🧠 Hippocampus

A living memory system for OpenClaw agents with importance scoring, time-based decay, and automatic reinforcement—just like a real brain.

The Concept

The hippocampus runs in the background, just like the real organ in your brain.

Your main agent is busy having conversations—it can't constantly stop to decide what to remember. That's what the hippocampus does. It operates as a separate process:

1. Background encoding: A cron job extracts signals, scores them, and uses LLM to create concise summaries
2. Automatic decay: Unused memories fade over time (daily cron)
3. Semantic reinforcement: When similar topics come up again, existing memories strengthen automatically

The main agent doesn't "think about" memory—it just recalls what it needs, and the hippocampus handles the rest.

Features

• Importance Scoring: Memories rated 0.0-1.0 based on signal type
• Time-Based Decay: Unused memories fade (0.99^days)
• Semantic Reinforcement: LLM detects similar topics → reinforces existing memories
• LLM Summarization: Raw messages → concise facts (via sub-agent)
• Fresh Install Friendly: Defaults to last 100 signals (not entire history)
• OpenClaw Integration: Bridges with memory_search via HIPPOCAMPUS_CORE.md

Installation

cd ~/.openclaw/workspace/skills/hippocampus
./install.sh                    # Basic (last 100 signals)
./install.sh --with-cron        # With encoding + decay cron jobs
./install.sh --signals 50       # Custom signal limit
./install.sh --whole            # Process entire history

Or via ClawdHub:

clawdhub install hippocampus

Quick Usage

# Load core memories at session start
./scripts/load-core.sh

# Search with importance weighting
./scripts/recall.sh "project deadline"

# Run encoding (usually via cron)
./scripts/encode-pipeline.sh

# Apply decay (usually via cron)
./scripts/decay.sh

Brain Dashboard

Visual dashboard showing all installed brain skills.

Access the Dashboard

Option 1: Auto-generated on install

./install.sh  # Creates brain-dashboard.html automatically

Option 2: Generate manually

./scripts/generate-dashboard.sh

Option 3: Open in browser

# macOS
open ~/.openclaw/workspace/brain-dashboard.html

# Linux
xdg-open ~/.openclaw/workspace/brain-dashboard.html

# Or open directly in browser:
# file:///home/USER/.openclaw/workspace/brain-dashboard.html

Features

• 🧠 Memory tab with stats and top memories
• 🎭 Emotions tab (if amygdala installed, or install prompt)
• ⭐ Drive tab (if VTA installed, or install prompt)
• Reads avatar/name from IDENTITY.md
• Auto-regenerates on every cron run (stays fresh)

How It Works

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│ Preprocess  │────▶│   Score &   │────▶│   LLM       │
│  signals    │     │   Filter    │     │  Summarize  │
└─────────────┘     └─────────────┘     └──────┬──────┘
                                               │
         ┌─────────────────────────────────────┘
         │
         ▼
┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│  Semantic   │     │   Store in  │     │    Decay    │
│  Reinforce  │────▶│  index.json │◀────│ (0.99^days) │
│  OR Create  │     │             │     │             │
└─────────────┘     └─────────────┘     └─────────────┘

Memory Domains

Domain	Contents
user/	Facts about the human
self/	Agent identity & growth
relationship/	Shared context & trust
world/	External knowledge

Decay Timeline

Days Unused	Retention
7	93%
30	74%
90	40%

Cron Jobs

The encoding cron does:

1. Extract signals from conversation
2. Score by importance
3. LLM compares to existing memories (semantic matching)
4. Similar topic → Reinforce existing memory
5. New topic → Create concise summary

# Encoding every 3 hours
openclaw cron add --name hippocampus-encoding \
  --cron "0 0,3,6,9,12,15,18,21 * * *" \
  --session isolated \
  --agent-turn "Run hippocampus encoding..."

# Daily decay at 3 AM
openclaw cron add --name hippocampus-decay \
  --cron "0 3 * * *" \
  --session isolated \
  --agent-turn "Run decay.sh..."

Requirements

• Python 3
• jq
• OpenClaw

AI Brain Series

Building cognitive architecture for AI agents:

Part	Function	Status
hippocampus	Memory formation, decay, reinforcement	✅ Live
amygdala-memory	Emotional processing	✅ Live
vta-memory	Reward and motivation	✅ Live
basal-ganglia-memory	Habit formation	🚧 Coming
anterior-cingulate-memory	Conflict detection	🚧 Coming
insula-memory	Internal state awareness	🚧 Coming

Based On

Stanford Generative Agents: "Interactive Simulacra of Human Behavior" (Park et al., 2023)

License

MIT

---

Memory is identity. Text > Brain. Part of the AI Brain series.