Now Available — PolyDocs v0.3 (93 Components, 73 Profiles)

MCP Servers that
Actually Ship

Production-grade MCP products by Arcflame Interactive. PolyDocs is a programmable document rendering engine — your agent composes components, we handle the format and layout.

Explore PolyDocs GitHub
93
Semantic Components
73
Document Profiles
10
Design Themes
5
Output Formats
Products

Our MCP Portfolio

Production-grade MCP servers designed for real AI agent workflows. Each product is self-contained, documented, and ready to deploy.

Arcflame PolyDocs — Document Rendering Engine for AI Agents

Arcflame PolyDocs

v0.3.0
Rendering Engine Semantic AI-Powered

A programmable document rendering engine for AI agents. Compose professional documents using high-level semantic components — not paragraphs or shapes. 93 components, 73 profiles, 10 themes.

Production Ready Learn More
Arcflame Archon — Architecture Governance for AI Agents

Arcflame Archon

v0.1.0
Governance Design Review Challenge Mode

An architecture governance engine for AI agents. Multi-lens reviews for APIs, databases, plans, and designs. Challenge Mode actively tries to break your proposal — stage-gated from intent to completion, with pass/fail verdicts and engineering memory that learns from every decision.

Production Ready Learn More
Arcflame Cinema — AI Video Production for MCP

Arcflame Cinema

v1.0
AI Film Studio Video Generation ComfyUI

A complete AI film studio over MCP. Reference-driven clip generation, scene scripting, AI scoring and TTS voices, HD upscale, and one-call publishing. Direct a finished film — script to screen — straight from your agent.

Production Ready Learn More

Coming Soon

In Development
Preview

More MCP products are in active development. Follow our GitHub and this space to be the first to know about new releases.

Coming Soon Follow Updates
Understand

What is MCP?

The Model Context Protocol is the open standard that lets AI agents talk to real tools, APIs, and data sources.

Model Context Protocol (MCP) is an open protocol developed by Anthropic that standardises how AI applications connect with external tools and data sources. Think of it as USB-C for AI — a universal connector that any MCP-compatible client can use to extend what an LLM can do.

Instead of every AI tool building its own integration, MCP provides a standard interface. One server, any client. Build once, deploy everywhere.

🔌

Universal Compatibility

Works with Claude Desktop, opencode, Continue.dev, Cursor, Windsurf, and any MCP client.

🔒

Local-First Security

Runs as a local child process. Your data never leaves your machine unless you want it to.

⚡

Zero Infrastructure

Install via pip/npm, configure in JSON, and go. No servers, no daemons, no cloud dependencies.

opencode.json — ~/.config/opencode/opencode.jsonc 
{
  "mcp": {
    "arcflame-polydocs": {
      "type": "remote",
      "url": "https://polydocs.arcflameinteractive.com/mcp",
      "headers": {
        "Authorization": "Bearer <your arcflame token>"
      },
      "enabled": true
    }
  }
}
Why Arcflame

Built by Developers, for Agents

Every MCP server we ship is designed with a single principle: give AI agents real capability, not boilerplate.

Semantic, Not Syntax

Your agent says "executive summary" not "left-align 14pt bold." We handle the rendering — DOCX, PPTX, HTML, PDF — from a single semantic plan.

Production Grade

73 document profiles, 93 semantic components, 10 themes, 5 output formats, universal diagram support, and full MCP compliance.

Sandboxed by Default

Output path sandboxing, image URL allowlisting, HTML escaping, SSRF protection, and temp auto-cleanup. Secure by construction.

Cost Predictable

Open-source core. Self-hosted. No per-seat fees, no API credits, no vendor lock-in. You control your infrastructure and your costs.

Extensible by Design

Plugin system for components, export formats, themes, and asset providers. Add your own without forking. Entry-point and filesystem discovery.

AI-First Architecture

Intent-aware planning, audience detection, medium optimisation, and ComfyUI integration for AI-generated visuals. Built for the agent era.

Custom MCP

Want a custom MCP built for your applications?

Tell us what you need — we'll build and host a dedicated MCP server tailored to your stack.

Specify your requirements:

  • → What APIs or data sources should the MCP connect to?
  • → What tools / functions should the AI agent have?
  • → Which MCP clients will you use (opencode, Claude, Cursor, etc.)?
  • → Expected call volume and hosting preferences
Send Request to [email protected]
See How We Build — Custom MCP

Ready to Ship?

Give your AI agents real document-generation superpowers. Get your Cloudflare-gated license and connect in minutes.

Get Started with PolyDocs