RunFlow CLI

Command line interface to manage AI agents via API Portal. Create agents, manage knowledge bases, handle prompts — all from the terminal.

Key Features

• 🚀 Create agents from templates with rf create
• 🧠 Manage knowledge bases (vector stores) with rf kb
• 📝 Full prompts management with CRUD and editor integration
• 👥 Multi-tenant profiles - switch between API keys easily
• 🧪 Local testing with web interface and live reload

📦 Installation

npm i -g @runflow-ai/cli

⚡ Quick Start

# 1. Authenticate with your API key
rf login

# 2. Create a new agent from template
rf create

# 3. Navigate to the created folder
cd my-agent/

# 4. Test locally with web interface
rf test

# 5. Deploy to production
rf agents deploy

🛠️ Commands

rf create - Create New Agent ⭐

Interactive agent creation with templates from the API.

# Interactive mode
rf create

# Non-interactive mode (for scripts/AI)
rf create --name my-agent --template starter
rf create -n my-agent -t rag-agent --yes

Options:

• -n, --name <name> - Agent name (non-interactive)
• -t, --template <id> - Template ID or name (non-interactive)
• -y, --yes - Auto-install dependencies without prompting

Flow:

1. Enter agent name (or use --name)
2. Choose template (or use --template)
3. Agent created on server with repository
4. Repository cloned locally
5. Dependencies installed (auto with --yes)

Available Templates:

• starter - Minimal setup, perfect for beginners
• rag-agent - Agent with knowledge base integration
• webhook-handler - Process webhooks and integrations

rf login - Authentication

# Interactive login
rf login

# With API key directly
rf login --api-key sk-xxx...

# Save as named profile
rf login --profile client-acme

Options:

• --api-key <key> - API key to use
• --profile <name> - Save as named profile
• --api <url> - Custom API URL

rf switch - Switch Profile

Switch between saved profiles for multi-tenant support.

# Interactive selection
rf switch

# Direct switch
rf switch client-acme

rf profiles - Manage Profiles

# List all profiles
rf profiles

# Show current profile
rf profiles current

# Delete a profile
rf profiles delete client-acme

rf agents - Agent Management

rf agents list              # List agents (interactive menu)
rf agents get               # Show current agent details
rf agents clone             # Clone agent repository
rf agents pull              # Pull latest changes
rf agents deploy            # Deploy local changes
rf agents duplicate         # Duplicate agent
rf agents delete            # Delete agent
rf agents delete --yes      # Delete without confirmation

Options:

• -y, --yes - Skip confirmation prompts

Aliases: rf agent

rf prompts - Prompts Management

Full CRUD operations for prompt templates.

# Interactive mode
rf prompts

# List all prompts
rf prompts list

# Get prompt content
rf prompts get <name>

# Create new prompt (interactive)
rf prompts create <name>

# Create prompt (non-interactive - for scripts/AI)
rf prompts create my-prompt --content "You are a helpful assistant"
rf prompts create my-prompt --file ./prompt.txt

# Update prompt (opens editor)
rf prompts update <name>
rf prompts update <name> --content "New content"

# Delete prompt
rf prompts delete <name>
rf prompts delete <name> --yes  # Skip confirmation

# Render with variables
rf prompts render <name> '{"var": "value"}'

Options:

• -c, --content <text> - Prompt content (non-interactive)
• -f, --file <path> - Read content from file
• -y, --yes - Skip confirmation prompts

Aliases: rf prompt

rf kb - Knowledge Base Management

Manage vector stores for RAG (Retrieval Augmented Generation).

# List all knowledge bases
rf kb list

# Create new knowledge base (interactive)
rf kb create <name>

# Create KB (non-interactive - for scripts/AI)
rf kb create my-kb --embedding "OpenAI Small"
rf kb create my-kb -e openai-small

# Upload file
rf kb upload <kb-name> <file>

# Upload directory (with confirmation skip)
rf kb upload <kb-name> ./docs --yes

# Check processing status
rf kb status <kb-name>

# List documents
rf kb docs <kb-name>

# Remove document
rf kb remove-doc <kb-name> <doc-id>
rf kb remove-doc <kb-name> <doc-id> --yes  # Skip confirmation

# Semantic search
rf kb search <kb-name> "your query"

# Delete knowledge base
rf kb delete <kb-name>
rf kb delete <kb-name> --yes  # Skip confirmation

Options:

• -e, --embedding <config> - Embedding config ID or name (non-interactive)
• -y, --yes - Skip confirmation prompts

Supported Files: PDF, TXT, MD, DOCX

rf test - Local Testing

Start local development server with web interface.

cd my-agent/
rf test

# Custom port
rf test --port 4000

# Without opening browser
rf test --no-browser

Features:

• Zero config - auto-detects from .runflow/rf.json
• Web portal with real-time monitoring
• Live reload on file changes
• Traces saved to .runflow/traces.json

Other Commands

rf users list               # Manage users
rf credentials list         # List credentials
rf triggers list            # Manage triggers
rf executions list          # View execution history
rf health                   # Check API health

📖 Usage Examples

Complete Development Workflow

# 1. Login and create agent
rf login
rf create
# → Select "RAG Agent" template
# → Name: "support-bot"

# 2. Navigate and set up knowledge base
cd support-bot/
rf kb create support-docs
rf kb upload support-docs ./docs --recursive

# 3. Test locally
rf test

# 4. Deploy to production
rf agents deploy

Multi-Tenant Workflow

# Save profiles for different clients
rf login --profile acme-corp
rf login --profile tech-startup

# Switch between clients
rf switch acme-corp
rf agents list

rf switch tech-startup
rf agents list

Knowledge Base Management

# Create and populate KB
rf kb create product-docs
rf kb upload product-docs ./manual.pdf
rf kb upload product-docs ./faq.md

# Check status
rf kb status product-docs

# Test search
rf kb search product-docs "how to reset password"

# Manage documents
rf kb docs product-docs
rf kb remove-doc product-docs doc_123

Using with AI Tools / Scripts (Non-Interactive Mode)

All commands support non-interactive flags for automation with AI tools (Cursor, Copilot) or CI/CD:

# Create agent without interaction
rf create --name my-agent --template starter --yes

# Create prompt without interaction
rf prompts create support-bot --content "You are a support assistant"

# Create KB without interaction  
rf kb create docs-kb --embedding "OpenAI Small"

# Delete without confirmation
rf agents delete --yes
rf prompts delete old-prompt --yes
rf kb delete old-kb --yes

# Upload directory without confirmation
rf kb upload docs-kb ./documents --yes

Flag Reference:

| Command | Specific Flags | Common Flags | |---------|---------------|--------------| | rf create | -n, --name -t, --template | -y, --yes | | rf agents | - | -y, --yes | | rf prompts | -c, --content -f, --file | -y, --yes | | rf kb | -e, --embedding | -y, --yes |

All commands also support: --api-key and --api for authentication overrides.

🔧 Configuration

Environment Variables

RUNFLOW_API_URL=https://api.runflow.ai
RUNFLOW_API_KEY=sk-xxx...

Config File (~/.runflowrc)

{
  "currentProfile": "default",
  "profiles": {
    "default": {
      "apiKey": "sk-xxx...",
      "apiUrl": "https://api.runflow.ai",
      "tenantId": "tenant_123",
      "tenantName": "My Company"
    }
  }
}

Project Config (.runflow/rf.json)

Created automatically by rf create or rf agents clone:

{
  "agentId": "agent_uuid",
  "agentName": "my-agent",
  "tenantId": "tenant_123"
}

📝 License

MIT – (c) IFTL 2025