Provider Comparison: Bright Data, Oxylabs, Smartproxy, IPRoyal, Others
An honest survey of the major proxy providers in 2026. Pricing tiers, target fit, ergonomic quirks. Not sponsored.
What you’ll learn
- Identify the top-tier providers and their general strengths.
- Spot the differences between pricing models (pay-per-GB vs. per-IP).
- Pick a provider based on target and budget, not marketing.
There are 50+ proxy providers. The market consolidates around 5–7 majors. This lesson surveys them honestly, what each is best at, where they're weak. No affiliate links, no sponsorships, no rankings.
Numbers and capabilities reflect 2026 norms. Check current pricing pages before committing.
The major providers
| Provider | Founded | Position |
|---|---|---|
| Bright Data | 2014 | Largest, premium, expensive, deep enterprise features |
| Oxylabs | 2015 | Close competitor to Bright Data; SERP API strong |
| Smartproxy (now Decodo) | 2018 | Mid-market, good ergonomics, simpler pricing |
| IPRoyal | 2020 | Aggressively priced, growing fast |
| NetNut | 2017 | ISP-proxy specialist, fast static IPs |
| SOAX | 2019 | Mobile-proxy strong |
| Zyte (was Scrapinghub) | 2010 | Built around Scrapy; smart-proxy + scraping API hybrid |
What each is good at
Bright Data
The Walmart of proxies. Largest residential and mobile pools, complete coverage. Has a "Web Unlocker" product that handles fingerprinting + retries for you (premium pricing).
Strengths: enterprise contracts, compliance features, SOC2, geographic coverage anywhere.
Weaknesses: pricing complexity (50+ products, each priced separately), entry cost is high, dashboard is heavy.
When to pick: enterprise legal/compliance requirements; you need every region; you have budget.
Oxylabs
Close second to Bright Data on size. Specialty: their SERP Scraper API is one of the cleanest implementations on the market.
Strengths: high success rate on SERP, good documentation, responsive support.
Weaknesses: similar pricing complexity to Bright Data.
When to pick: SERP scraping is the main use case; you've tried Bright Data and want a similar tier with cleaner ergonomics.
Smartproxy / Decodo
Mid-market positioning. Simpler pricing, better UX, smaller pools than Bright Data but sufficient for most projects.
Strengths: simple checkout, clean dashboard, predictable bills.
Weaknesses: smaller mobile pool, less premium support.
When to pick: small to medium scraping operations; you want predictable monthly costs without enterprise sales calls.
IPRoyal
The price disruptor of the last few years. Aggressive pricing on residential, increasing in scale.
Strengths: low entry cost (you can start at $1.75/GB residential at the time of writing, verify currently), reasonable coverage.
Weaknesses: support is thinner; reliability has improved but historically variable.
When to pick: budget-constrained; you can tolerate occasional retry-and-move on.
NetNut
ISP-proxy specialist. They have direct ISP relationships, so their static residentials have excellent uptime.
Strengths: static IPs that stay alive; low latency.
Weaknesses: smaller rotating residential pool; premium pricing.
When to pick: session-based scraping with logged-in accounts; you need IPs that don't disappear mid-flow.
SOAX
Mobile-proxy strong. Granular targeting (carrier-specific, city-specific).
Strengths: mobile pool is sizable and price-competitive.
Weaknesses: residential coverage is more limited.
When to pick: mobile-app scraping; Instagram/TikTok/social.
Zyte
Originally Scrapinghub, the company behind Scrapy. Their "Smart Proxy Manager" auto-routes requests, handles retries, and rotates fingerprints, pay for outcomes, not bandwidth.
Strengths: tight Scrapy integration; "I just want the page" abstraction.
Weaknesses: per-request pricing can be expensive at scale; less granular control.
When to pick: you're already a Scrapy shop; you'd rather pay more for less ops.
Pricing models
Two dominant patterns:
Pay-per-GB
Residential and mobile usually. Bandwidth-metered. Hidden cost: any data downloaded (including failed retries, blocked pages, ads, CSS) counts.
Per-IP per month
Datacenter and ISP usually. Fixed IPs, unlimited bandwidth. Cheaper if you have stable workloads; more expensive if you only use them sporadically.
Per successful request
Some "scraping API" products (Zyte, ScraperAPI, ScrapingBee). Pay for delivered pages, not attempts. Simpler accounting; usually more expensive per-success but predictable.
The hidden costs
Costs that aren't on the pricing page:
- Failed retries. Every 429 retry burns bandwidth. Aggressive rate limiting can 2-3x your bill.
- Image/CSS downloads. Default HttpClient pulls all assets. Strip them out.
- Geographic targeting upcharge. Some providers charge premium for specific countries.
- Concurrency limits. Lower-tier plans cap concurrent sessions.
- Support quality. Premium tier provides response in hours; budget tier might be days.
How to evaluate
A practical evaluation runs against your real target:
- Spin up trials with 2-3 providers.
- Run an identical scrape (same URLs, same volume) against your real target.
- Measure: success rate, latency, bandwidth used per success.
- Compute effective cost per successful page across providers.
The provider with the lowest sticker price often has the highest effective cost when you factor retries.
Avoid
- Providers with no transparent pricing (require sales call for everything below $10k/month). Usually a sign of bad fit for small/medium scrapers.
- Providers that won't let you pause subscriptions. Some lock you in to annual commitments.
- Resellers. Wholesalers who buy from majors and resell, you're paying for an extra layer.
Pure scraping APIs vs raw proxies
A different category exists: scraping APIs (ScraperAPI, ScrapingBee, ScraperBox, Zyte API). You send them a URL; they handle proxies, fingerprinting, rendering, CAPTCHAs. Pay per successful page.
| Raw proxies | Scraping API | |
|---|---|---|
| Control | Full | Limited |
| Anti-bot handling | DIY | Provider does it |
| Cost per success | Lower (with effort) | Higher (predictable) |
| Time to first scrape | Days/weeks | Minutes |
For one-off projects or where engineering time costs more than the API premium, scraping APIs are pragmatic. Covered in §4.37.
Hands-on lab
For a target you care about:
- Sign up for free trials at 2 providers from different tiers.
- Run an identical 1,000-request scrape against your target.
- Record success rate and total bandwidth.
- Compute effective cost per success.
The numbers from your actual target are worth more than every blog comparison. Provider quality varies dramatically by site.
Quiz, check your understanding
Pass mark is 70%. Pick the best answer; you’ll see the explanation right after.