DataImpulse: the cheapest per-request proxy I measured, and the most honest

I clicked the red link on DataImpulse's pricing page (BLOCKED WEBSITES) expecting the usual landing page, a "coming soon," something to soften the list. It opened a public list of blocked sites with the KYC layers spelled out in front of it. No proxy company does that. In four years of scraping (SERPwatch at 2M SERP checks a month, a coupon system feeding 15+ sites, now my AI visibility tool), I've seen the walls buried in fine print, never published on the pricing page. So when I put six providers through the same 20-site probe set for proxy.report in September 2026, DataImpulse had already earned something before the first request.

Then the numbers came in, and they held: 76.8% success tied for best of the six, the only provider clean through c16, at $0.15/1k (the cheapest cost per successful piece of data I've measured). The price of entry: a 2,114ms median latency. You pay in time, not money.

• $0.15/1k: cheapest per-request cost in the market, ~5.8× oxylabs' $0.86
• 76.8% success across 20 real sites, tied for first with decodo
• The only provider clean through c16: 96.9% at c16 where everyone else tops at c8
• 148/148 unique IPs: 100% rotation, no repeats
• Zillow at 95%: best of all six, where oxylabs, webshare, byteful get 0%
• The catch: 2,114ms median latency, the price of entry

The only provider that held at c16

I ran the same 20-site probe set across six providers (same transport stack, same concurrency ramp, no retries), and DataImpulse came back at 76.8% success, tied for the top with decodo. That alone wouldn't have surprised me. What did: the concurrency profile. Everyone else tops out at c8; DataImpulse holds clean through c16.

The c16 spike is likely the remaining block sites happening to sit outside that sample, but the shape is still the cleanest ceiling of any provider I tested. The cheapest provider is the only one that doesn't crater under load.

The auto-tier story is where pool reputation shows up. Zillow (a PerimeterX wall that oxylabs, webshare, and byteful all get 0% on) clears at 95%, the best of all six. Target 87%, rescued by the browser tier. Reddit 93%. Indeed 66%, stronger than most. The best performers read like a wishlist: booking 98.5%, imdb 98.5%, ozbargain 98.5%, grabon 97.7%, amazon 96.9%, agoda 95.9%, bestbuy 93.8%. And the hard-wall list is the shortest I measured: tripadvisor at 0%, the same universal wall every provider hits, and nothing else.

The honest gaps on the tough sites: ebay 38% (Cloudflare, partial), slickdeals 19%, yelp 4%. No provider in my tests cracks those clean, so I don't hold it against DataImpulse, but it's not the provider for yelp.

You pay in time, not money

The median latency is 2,114ms; p95 at 5,080ms, max at 10,394ms. That's the second-slowest of the six, ahead of only shifter at 2,402ms, and roughly 1.8× decodo (1,191ms) and 2.2× byteful (961ms). RPS comes in at 0.87, second-lowest throughput. The latencies move around under load (1,686ms at c1, 2,971ms at c8, back down to 1,617ms at c16), but the median is never fast.

For my coupon system, one scrape a day across 15+ sites, the 2.1s median means every run takes longer and the whole day's crawl finishes later than it should. It's a tax you feel most on time-sensitive work.

This is the tradeoff, and it's a clean one: DataImpulse wins the cost math, decodo wins the clock.

If your pipeline is latency-sensitive, decodo at 1.19s with 0.1% timeouts is the call. If you're buying successful data at scale, the 2.1 seconds is the price you pay for the cheapest per-request cost in the market.

The dashboard that tells you the truth

The dashboard looks old school on first look, and it's a precision instrument. The exclude feature lets you filter out a country or an ASN, something I've never seen laid out openly like this. The geo-targeting goes deeper than any provider I've tested: country + state + city + ZIP + ASN in a single endpoint.

Most providers let you pick state, or city, or ASN; DataImpulse lets you stack all of them. For location-based ad monitoring and zip-specific use cases, that's the exact control you need. The honest catch: the more you customize, the fewer IPs you get, and the dashboard tells you the usage with those parameters counts at a x2 coefficient. It says so on the line, not in fine print.

The sticky session setting is the same honesty. You can set a rotation interval up to 2 hours, but the company states it outright: they strive for stickiness but cannot guarantee it, and 30 minutes is the recommended maximum. That's the first provider I've seen admit the real behavior of sticky sessions in the dashboard itself.

The one friction point: the ASN list shows numbers, not names. I wanted to route through Verizon or AT&T and had to look up the ASN numbers on the internet to match them. It's a UX gap in an otherwise unusually honest tool.

The $5 trial that's honest, and the crypto that just works

There's no free trial, and DataImpulse doesn't pretend there is. It calls it a paid trial ($5 for 5GB), which at $1/GB residential, no subscription, no expiry, is almost free in my book. It's the same honesty as the blocked-website list: no bait, no fine print.

Then there's the crypto path. I tried to pay via crypto and it actually works. Most providers hook you into a wallet flow; DataImpulse generates a one-time address and lets you pay with it. The transaction can be a bit slow, but the mode of payment is a fresh breeze of good business values in a market where every step usually demands your identity upfront.

That's the cheapest per-request math in the market: $0.15 per 1k requests against oxylabs' $0.86, shifter's $0.33, decodo's $0.63. The volume discounts are marketing I haven't verified (20% at 1TB+, $0.8/GB over 1TB, custom pricing at 5TB+), so I'd treat those as promises to test, not facts.

The pool does the talking

The pool drain test came back perfect: 148 requests, 148 unique IPs, 100% rotation, no repeat. At 2-5M scrapes a month, fresh IPs with no fingerprinting build-up is the hygiene I need.

Country and ASN diversity are still unmeasured: the ipinfo backfill got rate-limited during the run, so I won't quote a pool size. What I did measure: the global egress split on the default gateway is US 18.8%, NL 10.4%, GB 6.2%, BR 6.2%, VN 6.2%. Not US-majority, so set a country filter if you need US traffic. The geo-tag is respected where I tested it: 94% US egress match, 100% on GB and DE.

The zillow number is the whole pool story in one row: 95% clear where oxylabs, webshare, and byteful get 0%. Same browser tier, same sites: the difference is which ASNs the exit nodes sit on and how those networks are treated. Pool size is marketing; exit-node quality is what matters, and DataImpulse's exit nodes clear PerimeterX where most don't.

Verdict

DataImpulse is the best value proxy I tested: 76.8% success tied for the top of six providers, the only one clean through c16, and the cheapest per-request cost in the market at $0.15/1k. The catch is latency: a 2,114ms median that makes it the wrong choice for speed-sensitive work. The transparency (the public blocked-website list, the honest sticky-session warning, the x2 coefficient stated on the line) is the most honest business conduct I've seen in this market.

Alternatives

If the latency or the untested gaps don't fit, the honest alternatives from the same test set:

• Decodo: $0.63/1k, 1,191ms p50, 0.1% timeouts. For speed-sensitive work, this is the pick.
• Webshare: $0.51/1k, 1,242ms p50, fastest European gateways. Budget speed without DataImpulse's c16.
• Oxylabs: $0.86/1k, same success tier, premium price without a success premium.

What's still untested

• SERP tracking: not run yet. I'll check it soon and update the review.
• Web Scraping API: not benchmarked. The 76.8% and 2,114ms figures in this review are from the residential proxy probe, a different product. My rule holds: I only trust a scraping product after it survives my own targets.
• Support: not tested this session. Nothing to say about it either way.
• ASN/country pool diversity: pending ipinfo backfill.
• Sustained 5-min throughput, home-IP baseline, mobile/ISP/datacenter tiers: not run.