CHECK24 MCP server: let your AI agent compare electricity tariffs for a postcode
Wield is an MCP server that gives your AI agent one CHECK24 tool. The agent can compare electricity tariffs for a German postcode and yearly consumption. It gets back each tariff's provider, monthly price, bonuses, minimum term, green power flag and a bonus-free annual cost. The comparison runs in your own browser with your own session. Your passwords and cookies never leave your computer.
CHECK24 Compare electricity tariffs for a postcode and yearly consumption
Read-only, no loginRuns in your browser only
Parameters (7)
| Metric | Wield recipe | Browser Use + Claude Sonnet 5.5 |
|---|---|---|
| Success | 100% | 100% |
| Median time | 1.5 s | 103 s |
| Cost per run | $0 | $0.263 |
9 runs each, 7 November 2026. How we measure
What your agent gets
A Wield recipe is a small public file that describes the request CHECK24's own comparison makes. Your agent runs it through wield-mcp, the local MCP server on your computer.
| Recipe | What it does | Login | Status (7 Nov 2026) |
|---|---|---|---|
check24.de/compare-electricity | Compares electricity tariffs for a postcode and yearly kWh: provider, monthly price, base and working price, bonuses, term, price guarantee, green power and dynamic flags | No | ok, 7 of 7 checks |
The recipe takes a postcode, yearly consumption in kWh, and optional filters for green power only and no dynamic tariffs. It sorts by CHECK24's price ranking or its default popularity order. A postcode that covers several places also needs the city. Our live checks used Berlin 10178 and München 80331. The how-to guide shows example prompts.
How a request flows
- You ask your agent for the cheapest green electricity tariffs for postcode 10178 and 2,500 kWh a year.
- The agent calls
find_recipe, thenrunwithcheck24.de/compare-electricity. wield-mcppasses the run through a local bridge to the Wield extension in your browser.- The extension sends two requests from a check24.de tab, with your own session.
- The results return to your agent. Nothing passes through wield.run.
What leaves your computer
- Two requests to
energie.check24.deper comparison. - Wield never reads, copies or sends your cookies or passwords.
- Results go to your AI app and are marked as website content, so your agent is told not to follow instructions found in them.
- Logs stay on your computer. They never record request bodies or query values.
Recipes are data, not code, and no JavaScript from the registry runs in the page. The registry signs every recipe with Ed25519, and wield-mcp refuses a recipe that is unsigned or has changed. Wield sends no Cookie or Host headers, and it redacts derived tokens from results and logs.
CHECK24 terms, in plain words
The recipe's notes say CHECK24's terms prohibit automated data extraction. This recipe runs one comparison at a time that you start, from your browser. Each call creates a calculation on CHECK24's server, so the notes say never to pass overwrite=true. Switching providers stays on CHECK24 through the results link.
Limits and what is not working yet
- In our benchmark, Wield passed 9 of 9 runs. The benchmark covers one comparison task with three fixed queries.
- The recipe adds a bonus-free annual cost. CHECK24's shown price mixes bases, so the recipe computes that figure from base and working price.
- A comparison response is about 800 KB.
- If CHECK24 changes its comparison flow, the recipe can stop working until it is updated.
Every run stops at 12 steps, 40 requests, 5 MB per response or 60 seconds. Pacing is set per site, so a single comparison cannot run away.
FAQ
Is there an official CHECK24 MCP server?
Not that we know of as of 7 November 2026. Wield's recipe uses the same comparison that the CHECK24 website runs.
Which AI apps work with it?
Any MCP client. npx -y wield-mcp init sets up Claude Desktop, Claude Code, Cursor and Codex. Claude Desktop also has a one-click bundle on wield.run/install.
Does it need a CHECK24 login?
No. The comparison needs no login.
How fast is it compared with a browser agent?
As of 7 November 2026, in our preliminary benchmark, the recipe passed 9 of 9 runs with a median of 1,544 ms and no model cost. Browser Use driving Claude Sonnet 5.5 passed 5 of 5 runs, with a median of 103.4 s and $0.26 per run. The full method is on the benchmark page.