This is the all-in-one package that includes everything you need:

• @marktoflow/cli — Command-line interface and workflow runner
• @marktoflow/core — Workflow engine, parser, state management, plugin system
• @marktoflow/gui — Visual drag-and-drop workflow designer
• @marktoflow/integrations — 38 service integrations and AI agent adapters

Install

npm install -g marktoflow

Usage

marktoflow init                              # Initialize project
marktoflow run workflow.md                   # Run a workflow
marktoflow run workflow.md --agent copilot   # Use your GitHub Copilot subscription
marktoflow run workflow.md --agent claude    # Use your Claude subscription (CLI login)
marktoflow run workflow.md --agent codex     # Use your Codex subscription
marktoflow run workflow.md --agent ollama    # Use Ollama (free, local)
marktoflow run workflow.md --agent vllm      # Use local llama.cpp / VLLM
marktoflow gui                               # Launch visual editor
marktoflow connect gmail                     # Setup OAuth
marktoflow serve --port 3000                 # Start webhook server

Why marktoflow?

• Use your existing AI subscriptions — Copilot, Claude, Codex — no extra API keys needed
• Markdown-native — Workflows are .md files — readable, auditable, version-controlled
• AI agents with tool calling — Agentic loops where models decide which tools to invoke
• 38 integrations — Slack, GitHub, Jira, Gmail, Stripe, Google Sheets, and more
• Local LLMs too — Optional llama.cpp, VLLM, Ollama for air-gapped or offline use
• Secure by default — Self-hosted, no telemetry, your data never leaves
• Visual editor — Optional drag-and-drop GUI with real-time execution
• Parallel execution — Run multiple AI agents concurrently for faster results
• MCP support — Native Model Context Protocol integration
• Structured output — JSON mode and JSON Schema validation for reliable AI responses
• Enterprise ready — RBAC, audit logging, cost tracking, AES-256 credential encryption

AI Agent Tool Calling

Build agentic workflows where the AI decides which tools to call:

steps:
  - action: ai.chatWithTools
    inputs:
      messages:
        - role: user
          content: "Research {{ inputs.topic }} and summarize findings"
      tools:
        - type: function
          function:
            name: search
            description: Search the web
            parameters:
              type: object
              properties:
                query: { type: string }
              required: [query]
      maxTurns: 5

Works with OpenAI, local llama.cpp, VLLM, LM Studio, and any OpenAI-compatible endpoint.

Learn More

Full documentation, examples, and source code:

github.com/marktoflow/marktoflow

License

AGPL-3.0 — Free for personal and open source use. Commercial licensing available — contact [email protected].