Sharper context. Fewer tokens.
OpenWolf is open-source middleware for Claude Code. It remembers your preferences, catches repeated mistakes, and tracks every token. You change nothing about your workflow.
Most Claude Code users hit their limits faster than expected.
The reason: Claude re-reads files it already saw, forgets corrections between sessions, and scans entire directories when a summary would do. There is no built-in way to see where your tokens go or why they're being spent.
OpenWolf fixes the visibility problem. It tracks every token, remembers across sessions, and gives Claude the context it needs without redundant reads.
Everything works invisibly
OpenWolf hooks into Claude Code's lifecycle. No commands to remember. It just makes every session smarter.
Invisible Enforcement
You type claude and work normally. Hooks fire automatically, tracking tokens, updating project maps, enforcing learned preferences.
Token Intelligence
Every token is estimated, tracked, and justified. Anatomy descriptions prevent unnecessary file reads. Repeated reads are caught and flagged.
No API Calls
All hooks are pure Node.js file I/O running locally. No network requests, no external services during normal operation. Optional scheduled tasks use your existing Claude subscription.
Self-Learning
Cerebrum tracks your preferences, mistakes, and decisions. Bug memory prevents the same fix twice. The system gets smarter every session.
Design QC
Capture full-page sectioned screenshots with one command. Claude evaluates the design inline. No external services, no extra cost.
Reframe
Ask Claude to help pick a UI framework. Built-in knowledge base covers 12 component libraries, from shadcn/ui to Aceternity UI to DaisyUI.
Three steps. Then invisible.
Initialize
Run one command in any project. Creates .wolf/ directory, registers hooks, scans all files.
Work Normally
Just use claude as you always do. Hooks fire invisibly, tracking, learning, enforcing. You don't interact with any of it.
Get Smarter
Every session, OpenWolf learns preferences, logs bugs, prevents repeated mistakes. View everything on the real-time dashboard.
The .wolf/ directory
Every project gets a .wolf/ folder containing state, learning memory, and configuration. Markdown is the source of truth.
anatomy.md
File index with descriptions and token estimates. Prevents unnecessary full-file reads.
cerebrum.md
Learned preferences, conventions, Do-Not-Repeat mistakes. Gets smarter every session.
memory.md
Chronological action log. Every read, write, and decision recorded per session.
buglog.json
Bug encounter and resolution memory. Searchable. Prevents re-discovering the same fix.
hooks/
6 Node.js hooks that fire on every Claude action. Pure file I/O, no network, no AI calls.
config.json
All settings with sensible defaults. Token ratios, cron schedules, dashboard port, exclude patterns.
The enforcement layer
Six hooks fire on every Claude action. They warn but never block. Pure Node.js. No network, no AI, no extra cost.
Start saving tokens today
One command to install. One command to initialize. Then it's invisible.