hacker-news-mcp
Provides Hacker News tools with founder-research features for analyzing Show HN launches, Ask HN discussions, and extracting startup insights.
README
hacker-news-mcp
A TypeScript MCP server for Hacker News with a practical founder-research layer on top.
It still gives you the raw Hacker News primitives you expect, but it also helps with the things indie developers actually care about:
- finding
Show HNlaunches in a niche - mining
Ask HNthreads for pain points - summarizing long discussions quickly
- extracting product, launch, and monetization signals
- comparing multiple launches for patterns
This project uses the official Hacker News API plus Algolia's Hacker News Search API. There are no external AI APIs, no database, and no browser automation.
What This MCP Can Do
Base Hacker News tools
searchgetStorygetStoryWithCommentsgetStoriesgetCommentgetCommentsgetCommentTreegetUsergetUserSubmissions
Founder-research tools
findShowHNfindAskHNsummarizeThreadextractStartupInsightscompareLaunches
Tool Overview
findShowHN
Search Show HN posts by keyword with optional filters.
Inputs:
query: stringpage?: numberhitsPerPage?: numberminPoints?: numberminComments?: numberdateFrom?: stringdateTo?: string
Returns normalized results with:
idtitleurlauthorpointsnumCommentscreatedAthnUrltextPreview
findAskHN
Search Ask HN posts for pain points and demand signals.
Inputs are the same as findShowHN.
summarizeThread
Summarize a story and comment thread using deterministic heuristics.
Inputs:
storyId: numbermaxDepth?: numbermaxComments?: number
Returns:
- story metadata
- key discussion themes
- notable pros
- notable cons
- repeated questions
- feature requests
- sentiment snapshot
- top comments used
extractStartupInsights
Extract founder-relevant signals from either one thread or a query.
Inputs:
storyId?: numberquery?: stringsourceType?: "show" | "ask" | "story" | "comment"maxItems?: number
Returns:
targetUserspainPointsalternativesMentionedobjectionsmonetizationSignalslaunchFeedbackopportunitiesrisksrecommendedFollowupslowConfidenceAreas
compareLaunches
Compare multiple Show HN launches.
Inputs:
storyIds: number[]
Returns:
- launch-by-launch comparison
- title style patterns
- topic categories
- common praise
- common criticism
- likely traction drivers
- skipped IDs when some stories cannot be loaded
Example Prompts
- "Find Show HN launches about CRM tools for founders with at least 30 comments."
- "Search Ask HN for pain points around bookkeeping automation."
- "Summarize HN thread
43130155withmaxDepth2 andmaxComments40." - "Extract startup insights from story
43130155." - "Analyze Ask HN results for
manual invoicingand return pain points plus monetization signals." - "Compare Show HN launches
43130155,42302015, and41806551."
Local Development
Requirements
- Node.js 18+
- npm
Install
npm install
Build
npm run build
Test
npm test
Run locally
node dist/index.js
Connect to MCP Clients
Build the server first:
npm install
npm run build
Then point your MCP client at the compiled entrypoint:
/absolute/path/to/hacker-news-mcp/dist/index.js
Replace /absolute/path/to/... with the real path on your machine.
Codex
If your Codex setup supports local stdio MCP servers, add an entry that runs:
{
"mcpServers": {
"hacker-news-mcp": {
"command": "node",
"args": ["/absolute/path/to/hacker-news-mcp/dist/index.js"]
}
}
}
This is the same pattern used by other local MCP clients: Codex launches the server process for you, so you do not need to manually keep node dist/index.js running in another terminal.
Antigravity
If you use Antigravity as an MCP client, configure hacker-news-mcp as a local stdio server with the same command:
{
"mcpServers": {
"hacker-news-mcp": {
"command": "node",
"args": ["/absolute/path/to/hacker-news-mcp/dist/index.js"]
}
}
}
If Antigravity expects an explicit transport type, use:
{
"mcpServers": {
"hacker-news-mcp": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/hacker-news-mcp/dist/index.js"]
}
}
}
Claude Desktop
Add this to your MCP config:
{
"mcpServers": {
"hacker-news-mcp": {
"command": "node",
"args": ["/absolute/path/to/hacker-news-mcp/dist/index.js"]
}
}
}
Cursor or other clients that expect type: "stdio"
{
"mcpServers": {
"hacker-news-mcp": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/hacker-news-mcp/dist/index.js"]
}
}
}
Install via npx
If you publish the package later, the config can become:
{
"mcpServers": {
"hacker-news-mcp": {
"command": "npx",
"args": ["-y", "@devabdultech/hacker-news-mcp"]
}
}
}
Generic stdio MCP clients
This server is intentionally simple:
- transport:
stdio - runtime:
node - entrypoint:
dist/index.js
If a client supports local MCP servers, you usually only need:
command: "node"args: ["/absolute/path/to/hacker-news-mcp/dist/index.js"]
After changing the config:
- Save the client configuration
- Restart or reload the client
- Ask the client to list tools or call
findShowHN
You only need to rebuild the server when the code changes. You do not need to manually run the server in the background for normal client usage.
Publish This Repo to GitHub for Free
If you have never published a project before, this is the straightforward path.
1. Make sure the repo is clean
Run:
npm run build
npm test
2. Create a new empty GitHub repository
On GitHub:
- Click
New repository - Name it
hacker-news-mcp - Leave it empty
- Do not add a README,
.gitignore, or license there
3. Initialize git locally
If this repo is not initialized yet:
git init -b main
git add .
git commit -m "Initial commit"
4. Connect the GitHub remote
Replace YOUR_USERNAME below:
git remote add origin https://github.com/YOUR_USERNAME/hacker-news-mcp.git
git push -u origin main
If GitHub asks for authentication, use GitHub Desktop, the GitHub CLI, or a personal access token in your normal setup.
Optional: Publish to npm Later
GitHub is free and easy.
npm publishing is separate.
Before publishing to npm, make sure:
- the package name is available
- your
npmaccount can publish the chosen scope - the README examples match the final package name
Current package name:
@devabdultech/hacker-news-mcp
If that scope is not the one you want long-term, change it before publishing to npm.
Risks and Gotchas
1. Git history reset
If you delete .git and reinitialize, you lose:
- previous commits
- original remote links
- old branch history
That is fine if you want a fresh public repo, but it is not reversible unless you backed up the old .git directory first.
2. Secrets
Before publishing, check that you are not committing:
.envfiles- API keys
- tokens
- private config files
- local logs
This repo currently ignores common env files and .omx/, which is good.
3. Client config paths
If your MCP clients still point at the old folder path, they will fail.
Use:
/absolute/path/to/hacker-news-mcp/dist/index.js
4. npm package name mismatch
GitHub repo names and npm package names do not have to match, but confusion there is a classic self-own.
If you publish to npm later, keep the README install command aligned with the real package name.
Project Structure
src/
api/ Hacker News and Algolia API clients
analysis/ Normalization, heuristics, and founder-research logic
models/ Lightweight domain models
schemas/ Zod request schemas
services/ Thin orchestration layer for higher-level tools
index.ts MCP server registration and tool handlers
tests/ Basic analysis tests
dist/ Compiled output
License
MIT. See LICENSE.
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。