Internal engineering reference for m9m — the n8n alternative without the bugs. m9m is a drop-in n8n alternative in Go: 5–10× faster, 70% lower memory, deterministic execution, zero npm dependencies.
For the public documentation site (installation, tutorials, node reference, API), see docs.neullabs.com/m9m — source under documentation/.
Official launch path: single binary distribution via package managers and release artifacts. Kubernetes content in this directory should be treated as experimental reference material.
- MCP Overview - Use Claude Code to orchestrate workflows conversationally
- 37 MCP tools for workflow automation, debugging, and plugin development
- Local mode (SQLite/Postgres) and Cloud mode (remote API) support
- Technical Specification - System overview and architecture
- Performance Report - Benchmarks and optimization
- Architecture Overview - System design
- Distributed Architecture - Multi-node setup
- Cluster Implementation - Clustering details
- Hybrid Architecture - Deployment patterns
- Scalability Analysis - Scaling considerations
- API Compatibility - n8n API compatibility
- API Implementation - Endpoint documentation
- Licensing & API - API access and licensing
- Deployment Guide - Getting started with deployment
- Production Guide - Production-ready configuration
- Deployment Overview - Deployment strategies
- SDK Overview - Embedding m9m in your applications
- Go SDK (
pkg/m9m) - Native Go library - Python Package (
packages/m9m-cli/python) - ctypes-based Python library - Node.js Package (
packages/m9m-cli/nodejs) - N-API native addon
- Plugin System - Plugin development and usage
- Plugin Architecture - Internal plugin design
- Workflow Versions - Versioning workflows
- Variables & Environments - Configuration management
- Node Overview - Node system documentation
- Transform Nodes - Data transformation nodes
- Trigger Nodes - Event trigger nodes
- Monitoring - Metrics and observability
- Security Review - Security considerations
- Migration from n8n - Migration guide
- n8n Feature Comparison - Feature parity matrix
- Refactor Plan - Modularization roadmap and todo list
- Roadmap - Future development plans
- Contributing - How to contribute
- Startup Time: < 500ms vs 3s for n8n
- Memory Usage: 150MB vs 512MB for n8n
- Execution Speed: 5-10x faster workflow execution
- Container Size: 300MB vs 1.2GB for n8n
- Built-in Prometheus metrics
- OpenTelemetry distributed tracing
- Horizontal scaling with queue systems
- Git-based workflow versioning
m9m is released under the MIT License.