geizhals-mcp

An MCP server that lets an AI search Geizhals price
comparison data: products, best prices, price history, ratings and categories.
It wraps the Geizhals mobile app's API (api.geizhals.net/gh/v9) and returns the
important fields to the AI.

Highlights · Setup · Run · Add to Claude · Tools · Notes · Troubleshooting · Disclaimer · License

Highlights

• Best price across every shop
  Find a product and get its best price, full price range, star ratings and test-review count, aggregated over all Geizhals merchants instead of a single listing.

• Know when to buy
  Pull the full price history and the current price drops (Schnäppchen), sortable by biggest percent off, so a model can answer "is this cheap right now or should I wait".

• From keyword to category
  Search by free text or browse the complete Geizhals category tree (2200+ categories shipped with the server) to drill into exactly the right segment, then compare a shortlist side by side.

• Price a used listing
  match_geizhals turns a classifieds ad headline into the Geizhals products it could be (with a confidence score), and get_model_price_range gives the new-price range across every variant of a model — the two numbers a "is this second-hand offer a deal" answer needs.

Setup

Requires Python 3.9+.

pip install -r requirements.txt

Run

python main.py

The server starts over streamable HTTP and prints where it's listening:

Starting geizhals-mcp server on http://127.0.0.1:8000/mcp

Point your MCP client at that URL. Host, port and path live at the top of
main.py.

Add it to Claude (and other MCP clients)

The server speaks streamable HTTP, so most clients only need the URL it printed
on startup.

Claude Code

One command, no config file:

claude mcp add --transport http geizhals http://127.0.0.1:8000/mcp

Add --scope user to have it in every project instead of only the current one.
claude mcp list shows whether the connection came up, claude mcp remove geizhals takes it out again.

Cursor, Codex, VS Code and other JSON-config clients

Same URL, in the client's MCP config:

{
  "mcpServers": {
    "geizhals": {
      "type": "http",
      "url": "http://127.0.0.1:8000/mcp"
    }
  }
}

Claude Desktop

Custom connectors (Settings > Connectors) are dialled from Anthropic's
cloud, so they cannot reach a server bound to your own machine. Two ways
around it:

Bridge over stdio with mcp-remote
(needs Node.js). Open Settings > Developer > Edit
Config, add the geizhals entry to claude_desktop_config.json and restart
Claude Desktop:

{
  "mcpServers": {
    "geizhals": {
      "command": "cmd",
      "args": ["/c", "npx", "-y", "mcp-remote", "http://127.0.0.1:8000/mcp"]
    }
  }
}

On macOS/Linux drop the Windows wrapper: "command": "npx" with
"args": ["-y", "mcp-remote", "http://127.0.0.1:8000/mcp"].

Or expose the server (cloudflared, ngrok, or run it on a box with a public
hostname) and add that HTTPS URL under Settings > Connectors > Add custom
connector. That route also makes the tools available in claude.ai and the
mobile apps, not just the desktop client but the API it talks to is then
reachable by whoever finds the URL, so put auth in front of it.

To stop confirming every call, open Settings > Connectors > geizhals and set
the tools' dropdown on the right to Always allow.

Tools

search_geizhals(query, ...)

Search products by keyword. Filters for country (loc/hloc), category and
manufacturer, plus sorting and paging. Every hit carries its current
best_price, avg_price, offer_count, currency, cheapest shop,
rating_* and an available flag, so a plain price-or-rating question needs no
follow-up call. Hits with no live offers add alltime_price_min /
alltime_price_max / alltime_last_date, so a list of discontinued hardware is
priceable without a call per hit. Also returns facet aggregates (categories,
manufacturers) to refine with.

Note that sort="p" narrows the result set as well as ordering it — Geizhals
cannot price-order a product that has no price, so unavailable products leave
both results and total.

get_product(product_id)

Full detail of one product: current best price and offer count, name, category,
rating, images, offer link, all-time price range and test-review count. A
product with no live offers gets its last_known_price / last_known_date
filled in instead of a bare null.

get_price_history(product_id, days=31)

Window (window_min/max/avg/change_percent) and all-time (alltime_min/max)
summaries kept strictly apart, plus a [iso_date, price] series.
include_series=False and granularity="week" keep long windows cheap.

get_product_ratings(product_id)

Average rating, totals, per-star counts and the reviews link.

browse_category(category, ...)

List the products in a category by its code, with price bounds and sorting.
Hits use the same field names as search_geizhals, and the price_range
returned alongside them is the category's real one (the search facet's is not).

compare_products(product_ids)

Compare several products side by side, specs included — properties that don't
apply to a product are dropped rather than shown as empty. Discontinued
products carry their last known price.

get_deals(sort="percent", ...)

Current price drops: products whose best price just fell, with the percent and
amount off. Sort by biggest percent drop, price, latest, popularity or
top deals, and filter with min_discount_percent / max_price. limit is a
target number of matches, not a fetch size.

get_model_price_range(model, ...)

What a whole model costs right now across all its variants — min, median,
max and the five cheapest — instead of one specific board partner card.
Sibling models (4070 Ti, 4070 Super) are kept out, and the result is always
scoped to one category: an unscoped search for a pricey product is dominated by
cases and water blocks below it and prebuilt systems above it, so the category
is detected when you don't pass one and reported back as category_code.

match_geizhals(title, ...)

Resolve a free-form listing title ("Gigabyte RTX 5070 Windforce OC 12GB NEU OVP
mit Rechnung") to Geizhals products, each with a 0–1 confidence and its live
price. When the top candidates are too close to separate, or nothing scores
convincingly, a note says so rather than letting a coin-flip look like an
answer.

list_categories(query)

Browse Geizhals' category tree to find the category codes browse_category
takes. Matches English or German names ("graphics cards" and "grafikkarten"
both work). The full tree ships with the server in data/categories.json.

Notes

• data/categories.json is a snapshot of Geizhals' category tree (title,
  title_de, id, children). main.py loads it for list_categories, which
  matches queries against both the English and German names.
• coverage.json holds the catalogue counts shown in the badges above.
  scripts/update_coverage.py refreshes it from the Geizhals homepage and a
  daily GitHub Action commits any change.
• API details are documented in search_api.md.
• This uses Geizhals' internal mobile API, not an official one. Be nice to it.

Troubleshooting

403 invalid request_fingerprint in payload — the HMAC secret or
signing scheme in main.py (GH_HMAC_SECRET, request_fingerprint,
API_HOST) no longer matches what the live Geizhals app sends. This is
reverse-engineered from the app and hardcoded, so it can drift whenever
Geizhals rotates the secret or changes the fingerprint scheme.

400 Bad Request — Geizhals validates several params against fixed sets
and rejects anything else outright. The two that bite: pagesize must be one
of 1/5/10/30/100/300/1000, and price_history's days must be 31/91/183/365.
The tools snap rows and days to allowed values, so a 400 usually means
another param drifted; the response body names the offending field. Tools
surface it as {"error": "..."} rather than raising the raw HTTP error, so the
model sees what went wrong instead of an internal URL.

Disclaimer

This is an independent, unofficial project and is not affiliated with, endorsed
by, or connected to Geizhals. "Geizhals" and all related trademarks belong to
their respective owners.

It's published for educational and research purposes only. It talks to Geizhals'
internal API, which is not meant for public use and may break or change at any
time. You are responsible for how you use it: respect Geizhals' Terms of Service,
robots rules and applicable law, and don't hammer their servers.

License

MIT LICENSE. Provided "as is", without warranty.