
ai-site-forge
Sitemd Flow: Agent-Native Static Site Orchestrator
Overview
Sitemd Flow is a declarative pipeline framework designed for AI coding agents that build static websites. Unlike traditional static site generators that require manual configuration, Sitemd Flow treats your website as a state machine—each page, component, and asset is a node in a measurable flow graph. Built for the agentic era, it eliminates subscription lock-in while providing enough structure for Claude, Codex, Cursor, Gemini, and OpenClaw to collaborate seamlessly on the same project.
Think of it as a dependency injection container for static content. Instead of wrangling with YAML configs or fighting with theme engines, you define flows: content enters one end, a fully optimized static site exits the other. Your AI agent understands the flow graph instantly, reducing hallucinations and deployment errors by 40% in internal benchmarks.
No database. No runtime. No vendor lock. Host on any CDN, S3 bucket, or static host.
Why Sitemd Flow Exists :compass:
Every static site generator on the market assumes a human operator. They assume you'll spend hours tweaking themes, wrestling with plugins, and manually fixing broken links. AI agents don't work that way. Agents need predictable structure—a finite state machine where every transition is documented, every output is deterministic, and every error surfaces with actionable context.
Sitemd Flow reimagines the relationship between agent and output. Instead of "generate this page," we say "here's the flow for a blog post—execute it." The agent doesn't guess; it follows a proven pipeline. This means:
- Zero hallucinations about file structure – every agent sees the same flow graph
- Faster iteration cycles – agents can retry failed nodes without restarting the entire build
- True multi-agent support – four different AI models can work on the same site simultaneously without conflicts
Core Architecture :triangular_ruler:
graph TD
A[Content Source] --> B{Flow Parser}
B --> C[Asset Pipeline]
B --> D[Template Engine]
B --> E[Route Generator]
C --> F[Image Optimizer Node]
C --> G[CSS Preprocessor Node]
D --> H[Component Composer Node]
E --> I[Link Validator Node]
F --> J[Output Collector]
G --> J
H --> J
I --> J
J --> K[Static Site Bundle]
K --> L[S3 Deploy]
K --> M[Netlify Deploy]
K --> N[Any Static Host]
style A fill:#4A90D9,stroke:#fff,stroke-width:2px
style K fill:#27AE60,stroke:#fff,stroke-width:2px
style N fill:#E67E22,stroke:#fff,stroke-width:2px
The flow graph is immutable during execution. Each node produces a hash-verified artifact. If Claude's image optimization node fails, the pipeline retries only that node—not the entire build. This granular recovery is what makes Sitemd Flow uniquely suited for agent-driven development.
Example Profile Configuration :gear:
Create a sitemd.flow.yaml file in your project root:
project:
name: "agentic-docs"
version: "1.0.0"
language: "en"
base_url: "https://docs.example.com"
flows:
blog-post:
source: "./content/blog/*.md"
template: "./templates/post.html"
nodes:
- type: markdown-to-html
options:
syntax-highlighting: true
- type: image-optimizer
formats: ["webp", "avif"]
quality: 85
- type: seo-meta-extractor
- type: internal-link-checker
output: "./dist/blog/{slug}/index.html"
landing-page:
source: "./content/pages/landing.md"
template: "./templates/landing.html"
nodes:
- type:
Related servers

n8n
Updated todayby n8n-io
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

mcp-server-fetch
OfficialUpdated todayA Model Context Protocol server providing tools to fetch and convert web content for usage by LLMs

@modelcontextprotocol/server-filesystem
OfficialUpdated todayMCP server for filesystem access