Otto.de MCP Server

Query otto.de products, prices, and reviews live from Claude, Cursor, n8n, or any MCP-compatible AI agent — without writing a single line of scraping code.

What it does

Runs as an MCP (Model Context Protocol) server on the Apify platform. Connect it once; your AI agent can then search otto.de products, browse categories, fetch full product details (GTIN, description, all images), and retrieve customer reviews — all in real time, as live tool calls.

Who uses this

• AI/agent developers — embed live otto.de product lookup into RAG pipelines, shopping assistants, and price-monitoring agents
• Claude / Cursor users — ask about otto.de products directly in your IDE or chat without leaving your workflow
• PropTech & SaaS developers — compose otto.de data into larger LLM workflows, n8n automations, or custom agents

MCP Tools

Tool	Description
search_otto_products	Keyword search — returns name, price, brand, rating, availability, image
browse_otto_category	Browse a category path (e.g. /moebel/sofas/)
get_otto_product	Full product detail for any URL — GTIN/EAN, SKU, description, all images, return policy
get_otto_reviews	Customer reviews for a product ID
get_otto_suggestions	Search autocomplete suggestions

Example tool call

search_otto_products(keywords="ecksofa grau", limit=5, sort="price_asc")

Returns:

{
    "query": "ecksofa grau",
    "count": 5,
    "results": [
        {
            "name": "VASAGLE Ecksofa L-Form 244 cm",
            "price": "339.99",
            "currency": "EUR",
            "brand": "VASAGLE",
            "availability": "InStock",
            "rating": "4",
            "url": "https://www.otto.de/p/vasagle-..."
        }
    ]
}

Connect to Claude

1. Deploy the actor on Apify in Standby mode.
2. Copy the Standby URL (format: https://<id>.apify.actor).
3. In Claude's MCP settings, add a remote server:

   URL: https://<id>.apify.actor/mcp
   Transport: Streamable HTTP
   

4. Ask Claude: "Search otto.de for a grey corner sofa under €400" — it calls the tool live.

Connect to Cursor / VS Code

Add to your MCP config:

{
    "mcpServers": {
        "otto-de": {
            "url": "https://<id>.apify.actor/mcp",
            "transport": "streamable-http"
        }
    }
}

One-shot scrape mode

Provide keywords in the input to run a classic one-shot scrape and push results to a dataset instead of starting the MCP server. Useful for scheduled batch jobs.

{ "keywords": "laptop 16 zoll", "limit": 100 }

Input parameters (one-shot mode)

Parameter	Type	Default	Description
keywords	string	—	Search query for one-shot run
limit	integer	50	Max results in one-shot mode

Scheduling & integrations

Schedule a daily one-shot run to keep a dataset fresh. Use webhooks to trigger downstream workflows in Make, Zapier, or n8n when the run completes.

FAQ

How is this different from the otto-scraper? This actor is an always-on MCP server — your AI agent calls it live, like an API. The otto-scraper is a batch tool that runs once and saves to a dataset.

Does get_otto_product handle Kasada anti-bot? Yes — it uses the Android app cookie trick to bypass Kasada on product detail pages without a browser.

Can I query any otto.de category? Yes — pass any category path like /computer/laptops/ or /sport-freizeit/fahrraeder/ to browse_otto_category.

What is the per-call cost? Each MCP tool call is charged as one Pay-Per-Event unit (see the Store listing for current prices). There is no monthly fee.