# 内容路由器 · 素材该做成什么

Analyze a piece of content across 11 dimensions and rank possible output formats, from slides to comics. Use it to choose a medium before starting the conversion work.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed · ★ 1
- Author: savantcat
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/savantcat/content-router
- Updated: 2026-10-21
- HTML: https://affpapa.org/en/nevibecode/mcp/item-e218e215

## What it does

The server builds a content profile using factors such as length, information density, structure, visual potential, depth, emotional tone, narrative tension, practical value, and timeliness. It then ranks suitable formats and gives scores with reasons behind each recommendation. Options include slide decks, infographics, short-video scripts, comics, newsletters, and reusable skills.

Run an analysis on a single file, process a directory in bulk, or import the engine from Python. The command `python engine.py batch ./我的素材库 --out 产出决策表.md` produces a decision table for a collection of source files. The `route()` method generates a dispatch prompt for selected formats and writes a JSON report. Long documents are split and summarized before profiling, and you can adjust profile dimensions before asking for new recommendations.

\`server.py` exposes five MCP tools over stdio or streamable HTTP. Install `mcp>=2.1` to run the server; no API key is needed. The tools perform calculations and read content without writing files or making external requests. The boundary matters: this project chooses a medium, such as an infographic or video, while rewriting an article for a particular social platform belongs to a separate workflow.

## Requirements
- Needs: Python and uv (commands run via uvx)

## Install

### Claude Code (native platform)

**Install**
- The MCP registry has no launch command for it: see the repository README

## When to use it
- Choose an output medium for a source article
- Build a format decision table for a content directory
- Compare recommendations and inspect their scoring reasons
- Generate a dispatch prompt for another agent
- Edit a content profile and recalculate format rankings

## Example prompts
- Analyze `README.md` and rank the best formats for this content.
- Process `./我的素材库` and save the decision table to `产出决策表.md`.
- Generate dispatch prompts to turn my article into a slide deck and an infographic.

## Similar
- [animated-infographic](https://affpapa.org/en/nevibecode/skills/animated-infographic.md) — skill · 1300 installs
- [youtube-research](https://affpapa.org/en/nevibecode/skills/youtube-research.md) — skill · 1100 installs
- [fanout](https://affpapa.org/en/nevibecode/mcp/houtini-ai-fanout.md) — MCP server · ★ 13
- [content-pattern-analyzer-sms](https://affpapa.org/en/nevibecode/skills/content-pattern-analyzer-sms.md) — skill · 1800 installs
- [baoyu-xhs-images](https://affpapa.org/en/nevibecode/skills/baoyu-xhs-images.md) — skill · 31300 installs
- [baoyu-infographic](https://affpapa.org/en/nevibecode/skills/baoyu-infographic.md) — skill · 32800 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md