Context7
Up-to-date docs for AI agents
作者提供的说明(英文原文)
Search willhaben.at for second-hand listings and rank the underpriced ones
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
WILLMEHR_MODE, WILLHABEN_COOKIE, WILLHABEN_MIN_INTERVAL_MS标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"willmehr": {
"command": "npx",
"args": [
"-y",
"willmehr"
],
"env": {
"WILLMEHR_MODE": "your-value",
"WILLHABEN_COOKIE": "your-value",
"WILLHABEN_MIN_INTERVAL_MS": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add willmehr --env WILLMEHR_MODE=your-value --env WILLHABEN_COOKIE=your-value --env WILLHABEN_MIN_INTERVAL_MS=your-value -- npx -y willmehr运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"willmehr": {
"command": "npx",
"args": [
"-y",
"willmehr"
],
"env": {
"WILLMEHR_MODE": "your-value",
"WILLHABEN_COOKIE": "your-value",
"WILLHABEN_MIN_INTERVAL_MS": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"willmehr": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"willmehr"
],
"env": {
"WILLMEHR_MODE": "your-value",
"WILLHABEN_COOKIE": "your-value",
"WILLHABEN_MIN_INTERVAL_MS": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
WILLMEHR_MODE envWILLHABEN_MIN_INTERVAL_MS env这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月7日
注册名称已通过 GitHub 账户 taneron 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
你需要填写:WILLHABEN_COOKIE
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(0.1.0)于2个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js。
同一分类中有:Context7, Codebase Memory, Exa
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Up-to-date docs for AI agents
Codebase knowledge graph for AI agents — 162 languages, sub-ms queries, 99% fewer tokens.
Connect AI agents to Exa for web search, content fetching, and multi-step research.
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
MCP servers for hunting deals on the German-speaking classifieds sites. TypeScript, over stdio locally or Streamable HTTP self-hosted; works with Claude Code, Claude Desktop, or any other MCP client.
Two servers ship from this repo, each registered separately:
| Server | Site | Entry point |
|---|---|---|
willmehr | willhaben.at (Austria) | dist/index.js |
kleinanzeigen-mcp | kleinanzeigen.de (Germany) | dist/kleinanzeigen/index.js |
They publish as two npm packages, willmehr and kleinanzeigen-mcp, released together from this repo at the same version.
The willhaben server was built by reverse-engineering the willhaben web app's own JSON API from a HAR capture. The Kleinanzeigen server is a port of jnslmk/kleinanzeigen-mcp — see Kleinanzeigen below.
claude mcp add willmehr -- npx -y willmehr
claude mcp add kleinanzeigen -- npx -y kleinanzeigen-mcp
Or from a clone, which is what you want for hacking on it or self-hosting:
git clone https://github.com/taneron/willmehr.git && cd willmehr
npm install && npm run build
claude mcp add willmehr -- node "$PWD/dist/index.js" claude mcp add kleinanzeigen -- node "$PWD/dist/kleinanzeigen/index.js"
Either way you get both servers locally, in full mode. Both are listed in the official MCP registry, as io.github.taneron/willmehr and io.github.taneron/kleinanzeigen, so a client that reads the registry can find them without this README. The willhaben account tools need a session cookie of your own — see Authentication; everything else works anonymously and needs no setup.
To run it as a shared HTTP server instead, on your own box, see Hosting. Nothing in this project phones home to anyone else's instance.
Both servers run in one of two modes. Same code, same binary — the mode is chosen at startup with --public / --full or WILLMEHR_MODE, and defaults to full.
full (default) | public | |
|---|---|---|
| willhaben account tools | registered | not registered |
.env / WILLHABEN_COOKIE | loaded | never read; ignored with a warning on stderr |
| willhaben request spacing | 400 ms | 900 ms |
| Kleinanzeigen spacing / pages / batch | 800 ms / 5 / 20 | 1500 ms / 3 / 10 |
public is what makes this safe to host for other people: the process holds no credential, and registers no tool that could use one, so a shared deployment cannot leak a session it never loads. The tighter pacing is the other half — in public mode every user shares one outbound IP, and both sites rate-limit per IP.
node dist/index.js --public # willhaben, anonymous
WILLMEHR_MODE=public node dist/kleinanzeigen/index.js
Account tools deliberately have no hosted equivalent. willhaben authenticates through Keycloak with PKCE, so there is no scoped token a third party could be granted — the only thing a hosted server could ask for is the raw session cookie, which reads messages and acts as the account. Anyone who wants those four tools runs full mode locally with their own cookie. See Hosting.
Four public tools, no account needed:
| Tool | Purpose |
|---|---|
willhaben_search | Search the marketplace with structured filters; auto-paginates |
willhaben_get_ad | Full detail for one ad: description, images, seller trust signals |
willhaben_find_deals | Price the market for a query, rank listings below the median |
willhaben_discover_filters | Resolve willhaben's opaque numeric category/brand ids |
Four account tools, requiring a session cookie:
| Tool | Purpose |
|---|---|
willhaben_my_profile | Your profile and numeric user id |
willhaben_my_watchlist | Ads saved to your watchlist |
willhaben_my_conversations | Recent buyer/seller chats |
willhaben_my_alert_count | Number of active saved-search alerts |
claude mcp add willmehr -- npx -y willmehr
Or from source:
npm install
npm run build npm run pack:kz # stage build/kleinanzeigen-mcp for publishing
willmehr publishes from the repo root. kleinanzeigen-mcp publishes from build/kleinanzeigen-mcp, staged by npm run pack:kz — it copies the built output the Kleinanzeigen entry point reaches (its own directory plus the shared mode module) and generates a manifest around it, so both packages ship the same commit at the same version. server.json and server.kleinanzeigen.json are the matching MCP registry entries.
Register it with Claude Code:
claude mcp add willmehr -- node "$PWD/dist/index.js"
Or in a client's MCP config:
{
"mcpServers": { "willmehr": { "command": "node", "args": ["/path/to/willmehr/dist/index.js"], "env": { "WILLHABEN_COOKIE": "BBX_JSESSIONID=...; x-bbx-csrf-token=..." } } } }
willhaben authenticates through Keycloak with PKCE, so there is no username/password grant to automate. Instead, lift the session from a logged-in browser:
npm run session -- ~/Downloads/www.willhaben.at.harnpm run check:authnpm run session writes .env (mode 0600, gitignored) and never prints cookie values. npm run check:auth exercises all four account tools and stops at the first 401 rather than repeatedly failing auth against your account.
The cookie must contain BBX_JSESSIONID (the session) and x-bbx-csrf-token. The CSRF token is read out of the cookie automatically — willhaben's web app mirrors that cookie into a same-named request header on every /webapi/ call, and the server reproduces that. WILLHABEN_CSRF_TOKEN only exists to override it.
The server loads .env from the package root itself, so MCP clients don't need to pass the cookie through their config.
BBX_JSESSIONID is sent with a five-day Max-Age, but the server invalidates it well before that — a session captured and replayed 38 minutes later was already rejected. It is not IP- or header-binding: the same jar, replayed verbatim with the browser's complete header set from the same machine and IP, still returns 401 while anonymous requests continue to work.
Treat the session as short-lived. Re-run npm run session right before you need the account tools, rather than expecting a capture to keep working for days.
| Variable | Default | Purpose |
|---|---|---|
WILLHABEN_COOKIE | — | Session cookie for account tools |
WILLHABEN_CSRF_TOKEN | from cookie | Override the CSRF token |
WILLHABEN_MIN_INTERVAL_MS | 400 (900 public) | Minimum spacing between requests |
WILLMEHR_MODE | full | full or public — see Modes |
Resolve ids before filtering. Categories and brands are opaque integers. Ask willhaben_discover_filters first — it returns each filter value with its id *and* a hit count for your current query:
Which laptop brands have listings under €300 in Vienna right now?
Narrow before ranking deals. willhaben_find_deals compares each listing against the median of its result set, so the query has to describe one kind of item. Searching thinkpad mixes €900 laptops with €5 docking stations and every "discount" it reports is noise. Searching thinkpad x1 carbon inside category 5831 (Notebooks) with priceFrom: 150 gives a median worth measuring against.
Watch for `suspiciouslyCheap`. Any listing under 20% of the median is flagged. Sometimes that's a genuine bargain; more often it's a placeholder price, a parts-only listing, or bait. The flag is a prompt to read the ad, not a verdict.
Sniping. Combine sort: "newest" with maxAgeHours: 2 to see only what appeared in the last couple of hours.
Find me underpriced ThinkPad X1 Carbons from private sellers in Vienna, posted in
the last day, and tell me which are worth messaging about.
The agent chains willhaben_discover_filters → willhaben_find_deals → willhaben_get_ad on the top candidates to check seller age and ad text.
Notes from the HAR analysis, since none of this is documented publicly:
/webapi/ route returns a bare 400 with an empty body without it. This looks exactly like a malformed query parameter and is the single easiest thing to get stuck on.navigatorGroups[], where each facet declares the query parameter it maps to and every legal value with a hit count. willhaben_discover_filters is a thin projection of that, which is why it stays correct when willhaben adds categories.1 newest, 3 price ascending, 4 price descending, 0 relevance.publicapi.willhaben.at and api.willhaben.at both refuse anonymous callers. The ad page is Next.js, so the full payload is read from __NEXT_DATA__, using the lighter /_next/data/ /...json route when the buildId is known and falling back to HTML when willhaben redeploys and the id rotates.src/normalize.ts flattens it.Requests are serialised through a single queue with a 400 ms floor between them, plus exponential backoff with jitter on 429/5xx. Please leave that in place.
Be aware that willhaben's robots.txt states *"It is expressively forbidden to use spiders, search robots or other automatic methods to access willhaben.at"*, disallows /webapi/, and specifically disallows the keyword, PRICE_FROM, PRICE_TO and periode parameters this server uses. Their Terms of Use likewise prohibit automated copying without consent. This tool performs exactly that kind of access.
That is a real constraint, not a formality: it's your account that would be suspended. This is built for personal, low-volume, occasional use against your own account. Don't point it at bulk collection, don't redistribute the data, and don't run it on a tight schedule.
aliildan/willhaben-mcp covers more verticals (cars, real estate, jobs) but is anonymous-only and has no price analysis. If you want broad read-only browsing rather than deal hunting on the marketplace, use that one instead.
A port of jnslmk/kleinanzeigen-mcp (Python) to TypeScript, covering kleinanzeigen.de, Germany's largest classifieds site. No account needed.
| Tool | Purpose |
|---|---|
kleinanzeigen_search | Search by keyword, location, radius and price; auto-paginates |
kleinanzeigen_search_by_url | Search from a pasted Kleinanzeigen URL, keeping filters the structured search can't express |
kleinanzeigen_get_listing | Full detail for one ad: description, images, seller, attributes |
kleinanzeigen_get_listings_batch | Full detail for several ids at once — the normal follow-up to a search |
kleinanzeigen_parse_search_url | Explain which filters a URL encodes. Makes no network request |
The intended flow is kleinanzeigen_search → pick interesting ids → kleinanzeigen_get_listings_batch. Search result
只显示了开头部分;其余内容请在仓库中阅读