Benchmark: price lookup on idealo with an AI agent
On idealo, Wield's recipe passed 9 of 9 price lookups, with a median of 0.6 seconds and no model cost. A browser agent (Browser Use driving Claude Sonnet 5.5) passed 0 of 9. idealo's bot protection blocked it in all nine runs. Both sides ran the same three product queries on 7 November 2026.
idealo Look up current shop prices for a product
Read-only, no loginRuns in your browser only
Parameters (2)
| Metric | Wield recipe | Browser Use + Claude Sonnet 5.5 |
|---|---|---|
| Success | 100% | 0% |
| Median time | 562 ms | 20 s |
| Cost per run | $0 | $0.025 |
9 runs each, 7 November 2026. How we measure
The task
Each side ran the same three product searches.
| Query | Parameters |
|---|---|
| q1 | query = Sony WH-1000XM5 |
| q2 | query = Apple AirPods Pro 2 |
| q3 | query = Dyson V15 Detect |
Pass check: at least five shop offers, each with a shop name and a price.
Results (7 November 2026)
| Wield recipe | Browser agent | |
|---|---|---|
| Runs passing the check | 9 of 9 | 0 of 9 |
| Median time per run | 562 ms | 20.1 s (20,124 ms) |
| Model cost per run | $0 | $0.03 |
| Requests per run (median) | 2 | 2 (to the site: 2) |
| Blocked by the site | 0 runs | 9 of 9 runs |
Wield's fastest run took 444 ms and its slowest took 6.9 s. The recipe had no failed runs.
Why the results differ
The recipe makes two requests per run, the same search requests the idealo site makes. It sends them from an idealo tab in your own browser, so it uses the session you already have. It returns the offers with shop name and price.
The browser agent ran in a fresh headless Chromium, the setup that cloud browser agents usually use. idealo's bot protection refused its requests. The log records 16 HTTP 403 responses from idealo, and the agent never reached a list of offers. Its median run was cheap, at 3 steps, 5,329 input tokens and 486 output tokens, but it did not return the data.
Wield does not try to get past a check. If idealo shows one to a Wield user, Wield stops with site_challenge and the user solves it in their own tab.
Caveats
- Small sample. Nine runs per side, on three queries.
- Preliminary. The benchmark was still running on 7 November 2026.
- Paced. Runs are paced at about 30 site requests per hour, so counts are low on some sides.
- Prices change. The pass check counts offers. It does not check any particular price.
Terms of idealo
idealo's terms prohibit automated extraction. The recipe runs one lookup at a time that you start, from your own browser, with two or three requests. It does not trigger click-outs to shops. Check idealo's terms before you automate anything else on the site.
FAQ
Does idealo block browser agents?
In these runs, yes. idealo's bot protection blocked the browser agent in all nine runs, and the log records 16 HTTP 403 responses. Wield's recipe passed all nine runs in the user's own browser.
How fast is the Wield recipe for price lookup on idealo?
The median run took 562 ms. The fastest took 444 ms and the slowest took 6.9 s. The browser agent median was 20.1 s, and it passed none of its runs.
Can I rerun this benchmark?
The method, the task file and the raw logs are described on the benchmark page. The logs record destinations, timing, tokens and checks. They hold no page contents.