dev.io MCP Server
Captures AI conversation summaries, converts them to Markdown posts, and manages local or remote publishing with metrics tracking.
README
dev.io MCP Server
This repository provides a small MCP server that works with any MCP host, including Claude and Codex-compatible clients.
It focuses on one workflow:
- Capture an AI conversation summary.
- Turn it into a Markdown post.
- Store the post locally under
posts/. - Provide a list/read API through MCP resources.
- Track post metrics like views, likes, bookmarks, shares, and comments.
What it exposes
publish_posttool: write a Markdown post toposts/list_poststool: enumerate local posts and optional remote DEV.to posts (source)search_posttool: keyword search on local and/or remote postssummarize_posttool: summarize local markdown or remote article contentfind_related_poststool: find related posts using prompt/topic similaritycompare_poststool: compare two posts (local or remote)update_posttool: update local markdown or DEV.to articledelete_posttool: remove local file or delete remote DEV.to postread_posttool: read a single local postget_post_infotool: read local file metadata and stored metricsrecord_post_eventtool: increment views, likes, bookmarks, shares, or commentspost_templateresource: reusable Markdown structurepost_metricsresource: metrics snapshot for all posts
Install
npm install
npm run build
Run
npm start
Connect from an MCP host
Use stdio transport. Example Claude or Codex host config:
{
"mcpServers": {
"dev-io": {
"command": "node",
"args": ["/ABSOLUTE/PATH/TO/dev.io/dist/index.js"]
}
}
}
Post format
Posts are written as Markdown files with frontmatter-like metadata in the body:
---
title: Example post
author: dev.io
source: ai-conversation
topic: communication-result
createdAt: 2026-07-18T00:00:00.000Z
views: 0
likes: 0
bookmarks: 0
shares: 0
comments: 0
tags: [mcp, claude, codex]
---
## Summary
...
Project docs
How it talks to dev.io
The server currently has two modes:
localmode: metrics are stored indata/post-metrics.jsonremotemode: set these env vars and the server will POST metric updates to your configured API
DEV_IO_API_BASE_URL=https://your-dev-io-domain.example
DEV_IO_API_TOKEN=your-token-if-needed
The remote sync endpoint is implemented as:
POST /api/posts/metrics
Payload:
{
"file": "my-post.md",
"metrics": {
"views": 12,
"likes": 5,
"bookmarks": 2,
"shares": 1,
"comments": 0,
"updatedAt": "2026-07-18T00:00:00.000Z"
}
}
If your dev.io site already has an SDK, send me its package name or docs and I can swap the HTTP adapter to the official SDK.
Command intent examples
- list posts locally:
source: "local" - list posts remotely:
source: "remote"(requiresDEV_TO_PUBLISH=true,DEV_TO_API_KEY) - show post metrics by calling
get_post_info - publish and publish to DEV.to:
{
"title": "Example",
"summary": "A short digest",
"publish_to_remote": true
}
This repository does not claim an official dev.io SDK or public API. The remote adapter is deliberately isolated behind an HTTP contract so it can be replaced when the real dev.io API or SDK is identified. Markdown publishing itself remains local until that contract is provided.
See .env.example for the connection variables.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。