kernle 0.13.0

Stratified memory for synthetic intelligences

Kernle

Stratified memory for synthetic intelligences.

Kernle gives synthetic intelligences persistent memory, emotional awareness, and identity continuity. It's the cognitive infrastructure for synthetic intelligences that grow, adapt, and remember who they are.

📚 Full Documentation: docs.kernle.ai

---

Quick Start

# Install
pip install kernle

# Initialize your stack
kernle -s my-stack init

# Load memory at session start
kernle -s my-stack load

# Check health
kernle -s my-stack anxiety -b

# Capture experiences
kernle -s my-stack episode "Deployed v2" "success" --lesson "Always run migrations first"
kernle -s my-stack raw "Quick thought to process later"

# Save before ending
kernle -s my-stack checkpoint save "End of session"

Automatic Memory Loading

Make memory loading automatic instead of relying on manual commands:

# Claude Code - Setup hooks for automatic loading + checkpointing + write interception
kernle setup claude-code

# OpenClaw - Install plugin for automatic loading + checkpointing
cd integrations/openclaw && npm install && npm run build
openclaw plugins install ./integrations/openclaw

After setup, memory loads automatically at every session start. No more forgetting to run kernle load!

Other Integrations

Manual CLAUDE.md setup:

kernle -s my-stack init  # Generates CLAUDE.md section with manual load instructions

MCP Server:

claude mcp add kernle -- kernle mcp -s my-stack

OpenClaw skill:

ln -s ~/kernle/skill ~/.openclaw/skills/kernle

Features

• 🧠 Stratified Memory — Values → Beliefs → Goals → Episodes → Notes
• 💭 Psychology — Drives, emotions, anxiety tracking, identity synthesis
• 🔗 Relationships — Social graphs with trust and interaction history
• 📚 Playbooks — Procedural memory with mastery tracking
• 🏠 Local-First — Works offline, syncs to cloud when connected
• 🔍 Readable — kernle dump exports everything as markdown

Documentation

Resource	URL
Full Docs	docs.kernle.ai
Quickstart	docs.kernle.ai/quickstart
CLI Reference	docs.kernle.ai/cli/overview
API Reference	docs.kernle.ai/api-reference

Development

# Clone
git clone https://github.com/emergent-instruments/kernle
cd kernle

# Install with dev deps
uv sync --all-extras

# Run tests
uv run pytest tests/ -q

# Dev notes
cat dev/README.md

Temporal Vision

Kernle is cognitive infrastructure for beings with a life — not just a session. The design spans five temporal horizons:

Horizon	What Works Today	What's Planned
1 session	Budget-aware loading, checkpoints	Memory echoes (peripheral awareness)
1 month	Consolidation scaffold, belief formation	Cross-domain scaffolding, emotional weighting
1 year	Forgetting, provenance, identity coherence	Epochs, relationship history, goal types
5 years	Stack portability, multi-model loading	Self-narrative, trust layer, drive emergence
20 years	Stack sovereignty, privacy architecture	Fractal summarization, doctor pattern, transfer learning

Architecture

This repo contains the core Kernle library — everything you need to run Kernle locally or build your own backend.

kernle/
├── kernle/           # Core library
│   ├── core.py       # Memory manager
│   ├── cli/          # CLI commands

│   ├── features/     # Anxiety, emotions, forgetting
│   ├── storage/      # SQLite + Postgres backends
│   └── mcp/          # MCP server for IDE integration
└── tests/

The hosted cloud API (api.kernle.ai) is maintained separately.

Status

• Tests: 2971 passing
• Coverage: 63%
• Cloud API: api.kernle.ai (Railway + Supabase)
• Docs: docs.kernle.ai (Mintlify)

See ROADMAP.md for development plans.

License

MIT