# 竞品内容 Pattern 雷达 · AI 答案里的重复说法

> 竞品内容 Pattern 雷达 · AI 答案里的重复说法 is an open-source MCP server for content publishing. Finds repeated claims, cited sources and sameness across competitor content for AI agents. It is available on API, CLI and self-hosted.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** API, CLI, self-hosted
- **Website:** https://savantcat.cn/mcp-radar.html
- **Repository:** https://github.com/savantcat/pattern-radar
- **MCP Registry name:** cn.savantcat/pattern-radar
- **Categories:** [content publishing](https://ai-powered.com/categories/content-publishing)
- **Listing updated:** 2026-11-18

## About

The README describes the server as a way to treat competitor pages or AI answers as a text corpus and count which phrases recur across them. It reports how often each phrase is covered, which sources the texts jointly lean on, and a similarity score for how alike the set is. The threshold for calling something a repeated pattern is calibrated by shuffling the corpus and taking the 95th percentile of the coincidental coverage, so the cutoff is derived from the data rather than fixed by hand.

The tool is deterministic statistics with no LLM-written interpretation, and it declines to report anything when the corpus has fewer than 12 items. The README states that it works on character n-grams without semantic clustering, so synonyms are not merged, and that it does not judge accuracy, plagiarism or whether a repeated claim is worth avoiding.

The site and README are written for content teams and agent builders who want to see what everyone in an industry is already saying before writing their own. Three read-only tools are listed: analyze_patterns, explain_method and self_check.

## Features

- **Cross-text pattern detection:** Finds phrases that recur across a batch of texts and reports their count and coverage ratio.
- **Shared source analysis:** Lists the sources that a group of texts commonly relies on.
- **Homogeneity score:** Produces a comparable number for how similar a set of texts is.
- **Self-calibrated thresholds:** Sets the repeat cutoff from a shuffled-corpus null distribution rather than a fixed constant.
- **Built-in self-check:** Runs control groups to confirm the measurement behaves as expected before results are trusted.
- **Minimum sample refusal:** Declines to report when the corpus has fewer than 12 texts.

## Good for

- Spot clichés that most competitors repeat before writing industry content
- Check whether a group of articles all draw on the same underlying sources
- Measure how similar a set of AI-generated answers to one question are
- Decide which claim to avoid or differentiate from in a content plan

## Tools

- `analyze_patterns`: Finds recurring patterns, frequently cited sources and homogeneity across a batch of texts.
- `explain_method`: Returns the definition of patterns, threshold method and tool limitations.
- `self_check`: Runs a control-group self-test to confirm the measurement works.

## FAQ

### How many texts does the analysis need?

The README says a corpus of fewer than 12 items is refused rather than given a score.

### Does it check whether repeated claims are true?

No. The README states it does not fact-check content and does not judge whether a repeated claim is correct or incorrect.

### Does it understand synonyms?

No. The README says it works at character n-gram level and does not merge synonyms or different phrasings of the same idea.

### Does it require installation or an API key?

The hosted endpoint is described as public, read-only and requiring no API key or installation. A local stdio or HTTP run from the repository is also documented.

### Is 竞品内容 Pattern 雷达 · AI 答案里的重复说法 free?

Yes. 竞品内容 Pattern 雷达 · AI 答案里的重复说法 is open source, so you can use and self-host it for free.

### What are the best alternatives to 竞品内容 Pattern 雷达 · AI 答案里的重复说法?

The closest alternatives to 竞品内容 Pattern 雷达 · AI 答案里的重复说法 are omniget, Awesome-MCP-ZH, Dev Feeds, Space Feeds and EarScribe MCP.

## Where this comes from

- Description: written from the project's README, the project's website and its MCP Registry entry.

## Alternatives

- [omniget](https://ai-powered.com/mcp/omniget): omniget is an open-source MCP server for ai models and llm tooling.
- [Awesome-MCP-ZH](https://ai-powered.com/mcp/awesome-mcp-zh): Awesome-MCP-ZH is an open-source MCP server for content publishing.
- [Dev Feeds](https://ai-powered.com/mcp/dev-feeds): Dev Feeds is a freemium MCP server for content publishing.
- [Space Feeds](https://ai-powered.com/mcp/space-feeds): Space Feeds is a freemium MCP server for content publishing.
- [EarScribe MCP](https://ai-powered.com/mcp/earscribe-mcp): EarScribe MCP is a free MCP server for design and media.
- [AI Dance MCP](https://ai-powered.com/mcp/ai-dance-mcp): AI Dance MCP is an open-source MCP server for content publishing.
- [Flux-2 Online MCP](https://ai-powered.com/mcp/flux-2-online-mcp): Flux-2 Online MCP is an open-source MCP server for design and media.
- [Food Feeds](https://ai-powered.com/mcp/food-feeds): Food Feeds is a freemium MCP server for content publishing.

---
Source: https://ai-powered.com/mcp/pattern-ai
