Works without tokens. Works offline. Works when everything else is broken.

Token-Free
Terminal Copilot

A deterministic CLI that installs, repairs, maintains, and operates your system using natural language — even when cloud LLMs are unavailable.

Install notokenView on GitHub
notoken
notoken> install claude
Installing Claude Code CLI...
✓ Claude Code CLI installed (v2.1.90)

notoken> install openclaw
Installing OpenClaw...
✓ OpenClaw installed and configured

notoken> install covex
Installing Covex...
✓ Covex installed and ready

notoken> the server seems sluggish
Plan:
  1. Check load averages
  2. Check memory usage
  3. Check disk space
  4. Find resource hogs

Why notoken?

I created this because when I run out of my Max Pro Plan, I have no way to give my system natural language commands without typing syntax-perfect commands the old-school way.

When the fancy AI is unavailable, this still gets useful work done.

What It Does

🔧

119 Built-In Intents

Services, Docker, Git, files, network, security, databases, logs, cron, packages, swap, permissions — all config-driven.

🚀

AI Tool Installer

Installs Claude Code, OpenClaw, Covex, Ollama, and other AI tools with a single command. No manual setup required.

🧠

Natural Language

Type 'restart nginx on prod' or 'check disk space' — no memorizing flags or syntax.

📊

Intelligent Analysis

Checks vCPU count vs load, flags full partitions, detects project types, assesses memory pressure.

🛡️

Safe by Default

Allowlists, confirmation gates, auto-backup before destructive ops, input sanitization, secret redaction.

🩹

Self-Healing Environment

Repairs broken setups — fixes OpenClaw, Node, npm, CLI tools, PATH issues, and dependency failures automatically.

🔄

Self-Healing Rules

Improves command understanding over time. Logs misclassified intents, proposes fixes, validates and applies new mappings.

📋

Playbooks

9 built-in recipes: health-check, security-audit, virus-scan, letsencrypt-setup, docker-check, and more.

🖥️

OS-Aware

Detects Ubuntu, Debian, Fedora, Arch, macOS, WSL, Windows. Adapts commands per distro and package manager.

📁

File Parsers

Reads passwd, shadow, .env, yaml, json, nginx, apache, BIND zone files. Knows where service configs live.

🔌

Ollama Auto-Detect

If Ollama is running locally, uses it automatically. Zero config. No cloud tokens burned.

Fixes Your System. Improves Itself.

Notoken has two layers of self-healing that work together — so it gets more capable the more you use it.

🩹
Layer 1

Self-Healing Environment

Repairs your system and installs. Detects and fixes broken state automatically.

  • ✓ Fixes broken OpenClaw, Node, npm installs
  • ✓ Repairs PATH and environment issues
  • ✓ Reinstalls corrupted dependencies
  • ✓ Recovers broken CLI tool configs
  • ✓ Auto-backup before destructive repairs
🔄
Layer 2

Self-Healing Rules

Improves how Notoken understands and executes commands over time.

  • + Learns from misclassified intents
  • + Updates command mappings automatically
  • + Improves NLP synonym matching
  • + Validates rule changes before applying
  • + Gets smarter with every interaction

Two Modes

Mode A — Offline

Deterministic. No LLM Required.

  • ✓ Installs & package management
  • ✓ System diagnostics & health checks
  • ✓ Service management (start/stop/restart)
  • ✓ Log inspection & error detection
  • ✓ File operations, archives, sync
  • ✓ Git & Docker lifecycle
  • ✓ Network tools (curl, dig, whois)
  • ✓ Security (firewall, fail2ban, certs)
  • ✓ Known repair flows
  • ✓ 9 built-in playbooks
Mode B — Enhanced

Optional LLM. When Available.

  • + Ambiguous phrasing interpretation
  • + Multi-step troubleshooting plans
  • + Deeper diagnostics
  • + Auto-learn new skills from failures
  • + Claude CLI, OpenAI API, or Ollama
  • + Explanation of complex errors
  • + Agentic workflows
  • + Falls back to offline mode gracefully

Get Started

Recommended

One-line install

# Download and run the installer
$ curl -fsSL https://notoken.sh/install | bash

Checks prerequisites, verifies integrity, and walks you through setup.

Quick start

Run with npx

# No install needed — run directly
$ npx notoken

Runs the latest version instantly. Requires Node.js 18+.

Usage

# Install globally
$ npm install -g notoken
# System health check
$ notoken doctor
# Install AI tools
$ notoken install claude
$ notoken install openclaw
$ notoken install covex
$ notoken install ollama
# One-shot command
$notoken "restart nginx on prod" --explain

Who Is This For?

→ Developers who run out of AI tokens
→ VPS owners who want guided sysadmin
→ Homelab users who want a smarter terminal
→ Field workers with unstable internet
→ Teams who want reproducible workflows
→ Anyone who wants a safety net for their systems

Consulting & Support

Available for hire. Whether you need a one-time setup or ongoing support, I can help.

🛠️

Server Setup

Full VPS/dedicated server provisioning. Nginx, Docker, CI/CD, DNS, SSL, mail, monitoring — production-ready from day one.

🤖

AI Tool Integration

Set up Claude Code, OpenClaw, Covex, Ollama, and custom AI workflows on your infrastructure. Local or cloud.

🔧

Custom Automation

Custom notoken playbooks, shell scripts, Docker pipelines, and deployment automation tailored to your stack.

🚨

Emergency Support

Server down? Broken deploy? Security incident? Get hands-on help fast. Available for urgent production issues.

👥

Team Onboarding

Get your team up to speed with notoken, AI-assisted DevOps workflows, and modern CLI tooling. Remote or on-site.

Dino Bartolome

Full-stack engineer & DevOps consultant. Creator of notoken.
Available for freelance, contract, and consulting engagements.

Hire MeGet SupportReport Issue