Olostep Task Review: Collecting Company Data from Websites

Olostep matched 334 of 343 expected details when we supplied exact page URLs and 158 of 343 when we supplied company homepages. Price matches dropped from 56 of 59 with exact URLs to 12 of 59 from homepages.

192 exact page URLs100 company homepagesVia treg

Results at a glance

Starting point Details found Runs with readable content Median time Time (P95) Confirmed API charge Charge per 100 details found
Exact page URL · Scrape 334/343 192/192 3.8 s 10.6 s $0.384 $0.11497
Company homepage · Crawl 158/343 79/100 26.4 s 89.9 s $1.64 $1.03797
What we checked Exact page URL Company homepage
Homepage titles 79/82 62/82
Product-page titles 73/75 21/75
Prices with billing units 56/59 12/59
Customer quotes 126/127 63/127

We matched returned text against 343 expected details from 82 companies. A run counted as readable when the scorer found nonempty text on an eligible page. Navigation can qualify as readable text even when a page's main content is missing. The charge per 100 details uses confirmed charges divided by matches. 11 homepage requests had unsettled or unavailable receipts. The confirmed charges and cost per 100 details are lower bounds for that test.

Test 1: Reading a page you already have

Result. On Handoff’s pricing page, Olostep returned an additional line that paired the Flex, Pro, and Scale discount amounts with their monthly units. All six prices matched in 3.1 seconds for $0.002. On Tiptap, three of six prices matched. The $59, $179 and $1,199 monthly checks did not match the returned text. Handoff pricing · Tiptap pricing.

All 9 misses had readable text from their reference page. The expected details did not match that text under the benchmark rules.

Setup. We passed one known page URL to olostep.web.scrape through treg, requested Markdown, and set max_age to zero.

Test 2: Starting from the company homepage

Result. From ZOKO’s homepage, the crawl returned ten pages, including the pricing page. All four prices matched in 39.1 seconds for $0.02. Dili was still in progress near the 90-second cutoff. A later status check showed it had completed with seven pages. The confirmed treg charge was $0.014. The Dili run earned no matches inside the window. ZOKO pricing · Dili.

Of the 185 missed checks, 181 had no readable text from their reference page in the response. The other 4 missed despite a readable reference page.

On Momentic, Scrape matched all seven prices on the supplied pricing URL. Crawl returned ten pages from the company homepage. The pricing page was absent, and the crawl matched none of those seven prices.

Setup. We submitted only the official homepage to olostep.web.crawl through treg with a ten-page limit, depth five, and robots.txt checking enabled in the request. The measured time covers job submission, polling, and result retrieval. Only completed crawls with results received within 90 seconds could score. All 100 cases contribute to the runtime and charges, and dispatch ran up to six cases at once.

When to use Olostep for this task

Use Scrape when you have the relevant product, pricing, or customer page. From a homepage, use Crawl to find candidate pages and check whether it reached the ones you need. Allow for jobs that are still in progress when you need the result.

Data and scoring

The November 17 tests through treg reuse the frozen November 15 URLs and reference checks. A hit is a normalized text match on an eligible URL. The pricing checks require the amount and billing unit. The November 15 Firecrawl and Context.dev tests ran through AgentRouter. The timings and charges here describe the November 17 workflows through treg, including their cache settings, concurrency, and retry handling.

Individual runs and scorecard lists inputs, returned URLs, checks, timings, and fees. Filter for provider olostep. Original comparison and methodology explains the shared dataset and scoring.