Verdict · Verified 18 August 2026
Thordata — priced well at the volumes research projects actually use
Residential from $0.65/GB at volume
Academic collection is bursty and mid-volume — a few hundred gigabytes across a fieldwork window rather than a sustained commercial pipeline — and that is the band where Thordata prices competitively: $1.50/GB at 50 GB and $1.00/GB at 150 GB against a pool of over 100 million residential IPs in 190-plus countries. The per-thousand SERP and scraper APIs are worth as much attention as the raw proxies for many study designs.
See Thordata pricing Opens on the vendor’s site · CASRAI referral link
Read the ethics and sizing guide first → — The technical decision is easy. The IRB, terms-of-service and data-protection questions are what stop projects — settle them before procurement.
Editorial disclosure: CASRAI has commercial referral arrangements with some of the vendors named on this page, and may earn a commission if you subscribe to them. We name them here regardless of whether a link is present. We only recommend tools our editorial team has independently researched. Read our full disclosure policy.
In summary
- Residential: $2.00 for 1 GB, $1.80/GB at 10 GB, $1.50/GB at 50 GB, $1.00/GB at 150 GB, $0.80/GB at 350 GB, $0.65–0.73/GB above 1,000 GB.
- Datacentre and static ISP $0.75/IP; mobile $2.20/GB. Verified 18 August 2026.
- APIs: SERP $0.70–0.80 per 1,000 responses, Web Scraper $0.50–1.00 per 1,000 results, Web Unlocker $1.00 per 1,000, Scraping Browser $2.50/GB.
- Pool: over 100 million residential IPs across 190+ countries.
- Pilot before committing to a package — bandwidth estimates are wrong by an order of magnitude more often than not.
Thordata pricing
Verified from the vendor pricing page, 18 August 2026
| Dimension | Product | Price | When it is the right choice |
|---|---|---|---|
| Residential | 100M+ IPs, 190+ countries | $2.00 (1 GB) → $1.50/GB (50 GB) → $1.00/GB (150 GB) → $0.65–0.73/GB (1,000 GB+) | Geographically-specific collection from sites that personalise or rate-limit |
| Datacentre | Server IPs | $0.75 per IP | High-volume collection from open or tolerant sources |
| Static ISP | Residential IPs, unlimited bandwidth | $0.75 per IP | Long sessions needing a stable identity |
| Mobile | Carrier IPs | $2.20/GB | Mobile-only content and apps |
| SERP API | Parsed search results | $0.70–0.80 per 1,000 | Cross-country search-results studies |
| Web Scraper API | Structured results | $0.50–1.00 per 1,000 | Common targets you would otherwise parse yourself |
| Web Unlocker | Challenge handling | $1.00 per 1,000 | Defended sites where raw proxies stall |
| Scraping Browser | Managed headless browser | $2.50/GB | JavaScript-heavy targets — watch the bandwidth |
Residential rates shown are the volume-tier prices; the vendor lists $0.65/GB as a sale rate against a $1.05/GB list. Check the current offer, since promotional pricing in this market moves frequently.
Does it price well at research volumes?
Proxy vendors optimise for commercial customers running continuous pipelines at terabyte scale. Academic collection looks different: a burst of activity during a fieldwork window, then months of nothing, at volumes measured in tens or low hundreds of gigabytes.
That matters because per-GB pricing is tiered, and the discounts that make headline rates attractive only arrive at volumes a research project will never reach. The honest comparison for a study is the mid-tier price, not the floor.
On Thordata’s published tiers a realistic project sits at $1.50/GB at 50 GB or $1.00/GB at 150 GB. For a collection exercise pulling, say, 80 GB across a three-month window, that is a budget line in the low hundreds of dollars — small enough to sit inside a consumables line rather than needing its own justification.
Per-IP products behave differently and suit different designs. Datacentre and static ISP at $0.75/IP bill for the address rather than the traffic, which favours sustained collection where you would otherwise pay for idle bandwidth. For a study that runs continuously over months against a tolerant source, a handful of datacentre IPs can be dramatically cheaper than per-GB residential.
The APIs deserve more attention than the proxies
Raw proxies give you IP addresses. You still write the collection code, handle challenges, parse markup, and repair everything when a target site redesigns — and that maintenance is usually the real cost of a scraping project, paid in researcher-hours that never appear in the budget.
For common targets a managed API removes that entirely. The SERP API at $0.70–0.80 per 1,000 responses is the standout for research use: studies of search-results variation across countries are a recurring design in computational social science and information-access research, and building a parser for a results page that changes layout every few months is a genuinely thankless job.
The Web Scraper API at $0.50–1.00 per 1,000 results covers other common structured targets. The Web Unlocker at $1.00 per 1,000 handles sites that actively challenge automated traffic.
The Scraping Browser at $2.50/GB is where budgets go wrong, so treat it carefully. A managed headless browser is necessary for JavaScript-heavy targets, but rendering pulls CSS, fonts, images, trackers and analytics alongside the content you want — per-page transfer can be several times higher than fetching HTML alone. Block asset loading where your research question does not need images, and pilot before committing.
Rule of thumb: if you need a well-defined structured dataset many people want, buy the API. If your targets are idiosyncratic — a specific set of institutional sites, a niche forum, a government register — you are writing custom code anyway and raw proxies are the cheaper input.
What to settle before any purchase
The technical choice is the easy part. These are what actually determine whether the project finishes, and they belong before procurement rather than after.
Ethics review. Publicly accessible does not mean exempt, and review boards increasingly say so — particularly where data concerns identifiable individuals or where aggregation enables re-identification. Approach your IRB or ethics committee with a written collection plan before you start; retrospective approval does not exist.
Terms of service and legal position. Most large platforms prohibit automated collection. The legal picture around scraping publicly-available data has moved in researchers’ favour in several jurisdictions but remains genuinely unsettled and varies by country. Involve your institution’s legal office for anything beyond small-scale public-page collection.
Data protection. Collecting personal data about EU or UK subjects engages GDPR regardless of whether the data was public — you need a lawful basis, and Article 14 transparency duties apply when you collect data about people from a source other than the people themselves. Your DPO should see the plan.
Pilot and measure. Buy the smallest package, collect a few hundred pages, and measure actual bytes per page against your real targets. Extrapolate from that. Buying a large bandwidth package on an unpiloted estimate is the most common way this budget line goes wrong, and researchers under-estimate by an order of magnitude routinely.
Full detail on all four is in the web scraping proxies guide, including the ethics section that no vendor page will write.
What to be careful about
Promotional pricing moves. The residential rate is listed as a sale against a higher list price, and this market discounts aggressively and often. Check the current rate rather than budgeting from this page, and get the figure that will actually be charged for your committed volume in writing before it goes in a grant.
Residential proxy sourcing is worth asking about. Residential IPs belong to real consumer connections, and how a provider obtains consent from those users varies across the industry and has been the subject of legitimate criticism. If your ethics committee asks where the IP addresses come from — and a good one might — you should have an answer from the vendor rather than a shrug. Ask directly during evaluation.
Support and documentation quality matter more than headline price for a research team without a dedicated engineer. Test the documentation during a small pilot.
Procurement fit. Prepaid per-GB credit does not always sit comfortably with institutional purchasing that expects invoices against a purchase order. Check with your finance office before assuming a card payment is acceptable on grant funds.
Pilot at the smallest tier before budgeting
Collect a few hundred pages from your real targets, measure the actual bytes per page, and extrapolate. That number is what belongs in the grant — not an estimate from a vendor calculator or from this page.
Residential from $0.65/GB at volume
See Thordata pricing Opens on the vendor’s site · CASRAI referral link
Frequently asked questions
How much does Thordata cost?
Residential proxies run $2.00 for 1 GB, $1.80/GB at 10 GB, $1.50/GB at 50 GB, $1.00/GB at 150 GB, $0.80/GB at 350 GB and $0.65–0.73/GB above 1,000 GB. Datacentre and static ISP are $0.75 per IP and mobile is $2.20/GB. Verified 18 August 2026.
Is Thordata good value for academic research?
At the volumes research projects actually use — tens to low hundreds of gigabytes — the relevant rates are $1.50/GB at 50 GB and $1.00/GB at 150 GB, which puts a typical collection exercise in the low hundreds of dollars. That is small enough to sit inside a consumables line rather than needing separate justification.
Should I buy proxies or a scraping API?
If you need a well-defined structured dataset that many people want — search results, common listings — buy the API, because the real cost of raw proxies is the researcher-hours spent maintaining parsers against sites that redesign. Thordata’s SERP API at $0.70–0.80 per 1,000 responses is the standout for cross-country search studies.
How much bandwidth will my project use?
More than you expect. A text-heavy HTML page is often 50–200 KB, so 100,000 pages is roughly 5–20 GB — but rendering JavaScript in a headless browser also pulls CSS, fonts, images and scripts and can multiply that several-fold. Always pilot a few hundred pages against your real targets and measure before buying a package.
Do I need ethics approval to use proxies for data collection?
The proxies are not what requires approval — the collection does. Review boards increasingly hold that publicly accessible does not mean exempt, particularly where data concerns identifiable individuals or aggregation enables re-identification. Approach your committee with a written plan before collecting, because retrospective approval is not available.
Where do residential proxy IP addresses come from?
Residential IPs belong to real consumer connections, and how providers obtain consent from those users varies across the industry and has attracted legitimate criticism. Ask the vendor directly during evaluation — if your ethics committee raises it, you want a documented answer rather than a shrug.
