notoken
v1.8.1
Published
Token-free terminal assistant — NLP-based server operations CLI
Maintainers
Readme
notoken
A token-free terminal assistant that installs, repairs, maintains, and operates your system — even when cloud LLMs are unavailable.
When the fancy AI is unavailable, this still gets useful work done.
notoken> check disk space
✔ server.check_disk...
── Analysis ──
⚠ CRITICAL: /mnt/c is 99% full (6.1G free)
✓ Root filesystem healthy (4% used)
notoken> install claude
Installing Claude Code CLI...
✓ Claude Code CLI installed (2.1.90)
notoken> the server seems sluggish
Plan:
1. Check server uptime and load averages
2. Check memory usage
3. Check disk space
4. Find resource hogsInstall
npm install -g notoken
notokenWhat It Does
Type server operations in plain English. 119 built-in intents — no tokens needed.
| Category | Commands | |----------|---------| | Services | restart, start, stop, enable, disable | | System | hostname, memory, disk, load, processes | | Docker | containers, compose up/down/build, cleanup, prune | | Git | status, log, diff, pull, push, branch, commit | | Files | find, grep, list, read, copy, move, archives | | Network | curl, dig, whois, ping, traceroute, ports | | Security | ufw, fail2ban, iptables, SSH keys, SSL certs | | Database | pg_dump, mysqldump, redis-cli | | Packages | install, uninstall, check |
Commands
notoken # Interactive mode (default)
notoken doctor # Diagnose + auto-fix system issues
notoken install <tool> # Install claude, openclaw, ollama, docker...
notoken uninstall <tool> # Clean removal
notoken setup openclaw # Guided OpenClaw + Matrix setup
notoken setup dev # Set up dev environment
notoken fix npm # Fix broken npm/git/docker/permissions
notoken check # Integration health check
notoken update # Check for updates and install
notoken logs <service> # Tail service logs
notoken "any natural language" # One-shot NLP commandTwo Modes
Mode A — Offline (no LLM required) Handles installs, diagnostics, maintenance, service ops, log inspection, repair flows, and 9 built-in playbooks. Works without internet.
Mode B — Enhanced (optional LLM) If Claude CLI, OpenAI, or Ollama is available, handles ambiguous phrasing, multi-step troubleshooting plans, and adaptive rules that improve over time.
# Enable adaptive rules
NOTOKEN_LLM_CLI=claude notoken --adapt
# Or use local Ollama (auto-detected, zero config)
ollama serve &
notokenInteractive Features
notoken> restart nginx on prod
notoken> do it again # repeats last command
notoken> same but on staging # changes environment
notoken> :adapt # toggle adaptive rules
notoken> :improve # run rule improvement now
notoken> :update # check for updates
notoken> :play health-check prod # run a playbook
notoken> :context # what the CLI remembers
notoken> :explain # toggle explain modeEnvironment Variables
NOTOKEN_LLM_CLI=claude # Claude CLI for LLM features
NOTOKEN_LLM_ENDPOINT=... # API endpoint (Claude/OpenAI)
NOTOKEN_LLM_API_KEY=... # API key
NOTOKEN_HOME=~/.notoken # Data directory (default: ~/.notoken)Support & Consulting
Need help with setup, custom workflows, or team onboarding?
- Environment setup — full dev/server machine configuration
- Custom automation — tailored playbooks and intent packs for your stack
- Team onboarding — standardized environments and training
- Troubleshooting — broken environment recovery, performance investigation
Contact: [email protected] | notoken.sh/consulting
Links
- Website: notoken.sh
- GitHub: offset25/notoken-cli
- Author: Dino Bartolome
License
MIT
