Neural Claude Code

Claude Code that learns and improves from every interaction.

---

The Problem

Every Claude Code session starts from zero. You explain your project structure, conventions, and preferences... again and again.

The Solution

Neural Claude Code remembers. It learns patterns, accumulates knowledge, and gets smarter over time.

Traditional: Execute → Forget → Repeat forever
Neural:      Execute → Learn  → Improve always

---

Quick Start

# 1. Clone
git clone https://github.com/brolag/neural-claude-code-plugin ~/Sites/neural-claude-code-plugin

# 2. Setup
cd ~/Sites/neural-claude-code-plugin && ./scripts/setup-hooks.sh

# 3. Add to shell
echo 'export CLAUDE_PLUGIN_ROOT="$HOME/Sites/neural-claude-code-plugin"' >> ~/.zshrc
source ~/.zshrc

Full Installation Guide →

---

Key Features

Feature	What it does
Self-Learning	Expertise files that grow smarter each session
Persistent Memory	Facts and patterns that survive restarts
Neural Loops	Autonomous coding sessions that iterate until done
Multi-AI	Route tasks to Claude, Codex, or Gemini
Research Swarm	Parallel agents for comprehensive research

---

Learn More

New here?	Know what you want?
Tutorials	How-to Guides
Step-by-step learning	Task-focused recipes

Need to look something up?	Want to understand?
Reference	Explanation
Commands, agents, config	Architecture, patterns

Full Documentation →

---

Quick Commands

# Remember something
/remember The API uses JWT tokens

# Run autonomous loop (unified command)
/loop "Fix all tests" --max 10

# Learn from any source (auto-detects type)
/learn https://github.com/user/repo

# Get multi-AI perspectives
/pv-mesh Should we use GraphQL?

# Parallel research
/research-swarm "React vs Vue vs Svelte"

All Commands →

---

Contributing

Contributions welcome! Open an issue or submit a PR.

---

License

MIT - see LICENSE

---

Built with Claude Code
_{Self-improving AI that learns from you}