MiniClaw CLI Documentation
The MiniClaw CLI provides a comprehensive set of commands for managing your AI agent infrastructure. All commands feature colorful, styled output with consistent design principles and enhanced user experience.
Table of Contents
- Installation and Setup
- Core Commands
- System Management
- AI Provider Management
- Content Management
- Communication Channels
- Job Scheduling
- Monitoring and Diagnostics
Installation and Setup
miniclaw install
Run the enhanced interactive setup wizard with Valkey installation.
miniclaw install
Features: - Step-by-step interactive setup process - Valkey automatic installation and configuration - Navigation between steps (back, forward, jump to specific steps) - Configuration review before applying changes - Workspace customization options - Multiple AI provider support (Ollama, OpenAI, OpenRouter) - Optional Telegram bot setup
miniclaw onboard
Alias for miniclaw install.
miniclaw onboard
Core Commands
miniclaw agent
Chat with the agent using styled output.
miniclaw agent -m "What is 2+2?"
miniclaw agent "Explain quantum computing"
echo "Summarize this" | miniclaw agent --stdin
Options:
- -m, --message MESSAGE: Chat message
- --stdin: Read message from stdin
- --json: Output raw JSON response
- --provider PROVIDER: Specify provider ID
miniclaw chat
Send chat messages with styled responses.
miniclaw chat "What's the weather like?"
miniclaw chat --source web "Tell me a joke"
Options:
- -m, --message MESSAGE: Chat message
- --stdin: Read message from stdin
- --json: Output raw JSON response
- --source SOURCE: Message source (default: cli)
- --provider PROVIDER: Specify provider ID
miniclaw gateway
Start the MiniClaw server with startup feedback.
miniclaw gateway
miniclaw gateway --host 0.0.0.0 --port 8080
Options:
- --host HOST: Bind host (default: 127.0.0.1)
- --port PORT: Bind port (default: 8787)
System Management
miniclaw status
Check system status with color-coded results.
miniclaw status
Displays: - Python environment information - Workspace status - Configuration file status - AI provider connectivity - Server status
miniclaw doctor
Comprehensive system diagnostics.
miniclaw doctor
Performs detailed checks including: - Python version verification - Workspace integrity - Configuration validation - Provider connectivity - Server health
miniclaw update
Update dependencies with progress indicators.
miniclaw update
Features: - Automatic dependency updates using uv or pip - Configuration defaults refresh - Progress feedback during updates - Success/failure indicators
miniclaw uninstall
Remove workspace with confirmation and preview.
miniclaw uninstall
miniclaw uninstall --yes # Skip confirmation
Options:
- --yes, -y: Skip confirmation prompt
AI Provider Management
miniclaw providers
Manage AI providers.
miniclaw providers list
miniclaw providers default --id ollama_default
miniclaw providers delete --id openai_fallback
Subcommands:
- list: List configured providers
- default --id ID: Set default provider
- delete --id ID: Delete a provider
- save: Create or update a provider
miniclaw providers save
Create or update an AI provider configuration.
miniclaw providers save \
--id openai_gpt4 \
--name "OpenAI GPT-4" \
--type openai_compatible \
--base-url https://api.openai.com/v1 \
--api-key sk-your-api-key \
--model gpt-4-turbo \
--temperature 0.7
Options:
- --id ID: Provider ID (required)
- --name NAME: Provider name (required)
- --type TYPE: Provider type (ollama, openai_compatible, openrouter, litellm)
- --base-url URL: Base URL for API
- --api-key KEY: API key for authentication
- --model MODEL: Default model name
- --temperature TEMP: Temperature setting (0.0-2.0)
- --timeout SECONDS: Request timeout
- --prompt-override TEXT: Override system prompt
- --disable: Disable provider
- --no-verify-tls: Disable TLS verification
Content Management
miniclaw skills
Manage skill files with CRUD operations.
miniclaw skills
miniclaw skills --json # Raw JSON output
miniclaw skills save
Create or update a skill file.
miniclaw skills save --id research_helper --file ./research_skill.md
echo "# Research Helper\nkeywords: research, analysis\n\nHelp with research tasks." | miniclaw skills save --id research_helper --stdin
Options:
- --id ID: Skill ID (required)
- --file FILE: Read content from file
- --stdin: Read content from stdin
miniclaw skills delete
Delete a skill file.
miniclaw skills delete --id research_helper
Options:
- --id ID: Skill ID to delete (required)
miniclaw memory
Manage memory files.
miniclaw memory list
miniclaw memory get --name soul.md
Subcommands:
- list: List memory files
- get --name NAME: Get memory file content
- save --name NAME: Save memory file content
miniclaw memory save
Save content to a memory file.
miniclaw memory save --name project.md --file ./project_notes.md
echo "# Project Notes\n\nKey decisions and constraints." | miniclaw memory save --name project.md --stdin
Options:
- --name NAME: Memory file name (required)
- --file FILE: Read content from file
- --stdin: Read content from stdin
Communication Channels
miniclaw telegram
Manage Telegram integration.
miniclaw telegram restart
miniclaw telegram test --chat-id 123456789 --message "Test message"
miniclaw telegram pairings
Subcommands:
- restart: Restart Telegram poller
- test --chat-id ID --message MSG: Send test message
- pairings: Get pairing status and requests
- pair-start [--ttl SECONDS]: Create pairing code
- pair-confirm --request-id ID: Confirm pairing request
- pair-reject --request-id ID: Reject pairing request
- unbind: Remove currently bound Telegram chat
Job Scheduling
miniclaw jobs
Manage scheduled jobs.
miniclaw jobs status
miniclaw jobs save
Create or update a scheduled job.
miniclaw jobs save \
--id daily_report \
--name "Daily Report" \
--prompt "Generate a daily summary of system activity" \
--interval 86400 \
--telegram-chat-id 123456789
Options:
- --id ID: Job ID (required)
- --name NAME: Job name (required)
- --prompt PROMPT: Agent prompt to run (required)
- --interval SECONDS: Run interval in seconds (min 10, default: 300)
- --disabled: Create job in disabled state
- --telegram-chat-id ID: Optional Telegram chat ID for job output
miniclaw jobs delete
Delete a scheduled job.
miniclaw jobs delete --id daily_report
Options:
- --id ID: Job ID to delete (required)
miniclaw jobs run
Trigger a job immediately.
miniclaw jobs run --id daily_report
Options:
- --id ID: Job ID to run (required)
Monitoring and Diagnostics
miniclaw health
Check server health status.
miniclaw health
miniclaw models
List available AI models.
miniclaw models
miniclaw models --provider ollama_default
Options:
- --provider PROVIDER: Provider ID (default provider if omitted)
miniclaw usage
Show token usage statistics.
miniclaw usage
miniclaw usage --limit 100
Options:
- --limit LIMIT: Max recent usage entries (default: 250)
miniclaw runtime
Display runtime information.
miniclaw runtime
miniclaw history
View chat history.
miniclaw history
miniclaw history --limit 50
Options:
- --limit LIMIT: Max history entries (default: 50)
miniclaw events
Monitor system events.
miniclaw events
miniclaw events --since-id 100 --limit 100
Options:
- --since-id ID: Start from event ID (default: 0)
- --limit LIMIT: Max events to fetch (default: 100)
miniclaw config
Manage configuration.
miniclaw config get
miniclaw config set --file ./new_config.json
miniclaw config raw-get
miniclaw config raw-set --file ./raw_config.json
Subcommands:
- get: Get current configuration
- set --file FILE: Update configuration from JSON file
- raw-get: Get raw configuration text
- raw-set --file FILE: Update raw configuration text
CLI Styling and User Experience
All CLI commands feature enhanced user experience with:
Color Coding
- Green ✓: Success operations
- Red ✗: Errors and failures
- Yellow ⚠: Warnings and cautions
- Blue ℹ: Informational messages
- Purple: Code snippets and technical terms
Visual Hierarchy
- Bold Headers: Section titles and important information
- Dimmed Text: Secondary information and metadata
- Structured Lists: Organized content presentation
- Progress Indicators: Feedback during long operations
Consistent Patterns
- Step-by-step Navigation: In setup and multi-stage operations
- Confirmation Prompts: For destructive operations
- Helpful Error Messages: With context and guidance
- Standardized Output: Consistent formatting across all commands
Examples
Complete Setup Workflow
# Install and configure MiniClaw
miniclaw install
# Check system status
miniclaw status
# Verify health
miniclaw health
# List available models
miniclaw models
# Start the server
miniclaw gateway
Content Management Workflow
# List current skills
miniclaw skills
# Create a new skill
echo "# Code Review
keywords: code, review, programming
Provide code review feedback focusing on:
- Best practices
- Security considerations
- Performance optimizations" | miniclaw skills save --id code_review --stdin
# Update memory
echo "# Project Context
- Using Python 3.11+
- Deployed on cloud infrastructure
- CI/CD with GitHub Actions" | miniclaw memory save --name project.md --stdin
# Verify changes
miniclaw skills
miniclaw memory list
Job Management Workflow
# Check current jobs
miniclaw jobs status
# Create a daily report job
miniclaw jobs save \
--id daily_report \
--name "Daily System Report" \
--prompt "Generate a daily report summarizing system activity, token usage, and any errors from the past 24 hours." \
--interval 86400
# Run immediately for testing
miniclaw jobs run --id daily_report
# Verify job status
miniclaw jobs status
Monitoring Workflow
# Check system health
miniclaw health
# View recent events
miniclaw events --limit 20
# Check token usage
miniclaw usage --limit 10
# View chat history
miniclaw history --limit 10
# Get runtime information
miniclaw runtime