The best TikTok scrapers, ranked

TikTok is the platform where the gap between what is technically easy and what is officially allowed is at its widest. The data is right there in a public feed, half a dozen products will hand it to you as JSON this afternoon, and the only sanctioned route is a research programme that begins by asking which university you work for. If the terms are what brought you here rather than the tools, they are quoted clause by clause in is scraping social media legal.

The official API, and why it is not on the list

Open developers.tiktok.com expecting a data platform and you get a games platform. The headline is about mini games, the nav runs Mini Games, Mini Dramas, Login and Identity, Share to TikTok, Research and Insights, Content Display. That ordering is the argument: TikTok's developer programme is built to put things into TikTok, and reading things out of it is a compliance obligation parked in fifth place.

The Research API behind that fifth item is genuinely good, and genuinely closed. Applicants must be affiliated with an eligible organisation, meaning academic institutions in the US, the EEA, the UK or Switzerland, not for profit research institutions in the EU under a beta, or academic and non profit organisations in Brazil working on youth safety. The research must be non commercial and pursue a public interest mission, and applicants must be "independent from commercial interests". You need a defined proposal, evidence of ethical review and a funding disclosure.

If you qualify for the Research API, stop reading and apply. Nothing below is as good, as cheap or as safe. If you do not qualify, no product below turns that no into a yes.

There is a second official surface worth knowing about, the Commercial Content API and the ad libraries built on it, which exist because of EU transparency rules and cover advertising rather than organic posts. If your question is what your competitors are running as paid, look there first, because it is free, sanctioned and nobody has to work around anything.

So what are the other nine standing on

TikTok's terms are shorter and blunter than Meta's. Among the prohibited activities is to "use automated scripts to collect information from or otherwise interact with the Services", and the terms open by stating that the services are provided "for private, non-commercial use". There is no logged out carve out to argue about and no equivalent of the litigation that shaped Meta's position. The category simply operates outside the terms and prices itself for that.

What that means in practice for a buyer is narrower than it sounds, and worth being unromantic about. The realistic consequences are blocking, an endpoint going quiet without warning, and the vendor absorbing an enforcement action you never see. The consequence that would land on you personally arrives only if a tool is doing something as you, from an account, which two entries below can be configured to do. That, rather than the terms in the abstract, is what the ranking weighs.

What ranked means here

• Does somebody else run the part that breaks? On TikTok that means proxies, request signing and a parser that survives the app changing shape.
• Can you price a job before running it? Per result and per credit can be multiplied. Per gigabyte cannot.
• How far is it from an account you care about? No account, somebody else's accounts, or yours are three different products sold under one word.
• Is the breakage visible? A published success rate, a status page and a changelog beat a feature list, and on TikTok they matter more than anywhere else.
• Can you leave with your data and without a rewrite?

1. Apify

The TikTok Scraper is maintained by Clockworks, an Apify affiliated developer, and on the day we checked it showed 243,331 total users, a 4.78 rating and pricing from $1.70 per 1,000 results with the platform usage included in that figure. Per result is the meter you want, because a job over three hundred hashtag videos is a multiplication you can do in your head before you spend anything.

The number worth staring at is on the same page and it is not the rating: 92.8% of runs succeeded. Set that beside the 100% the same marketplace reports for its first party Instagram actor and you have the most honest single statistic on this page. TikTok breaks more. Build the retry in from the start, expect a run to come back short sometimes, and treat any vendor whose marketing implies otherwise with the scepticism it has earned.

2. ScrapeCreators

If you are writing the code yourself, this is probably the one. A single x-api-key header, a TikTok API with profile, video, comment, reply, following and music endpoints, no account level rate limit, 100 free credits with up to 7,000 more claimable, credits that never expire, and cached results billed at nothing. Its front page claims 10 million requests a day at 98.2% success and a 3.12 second average.

It also has the most useful obscure endpoints in the category, including the TikTok song and sound scrapers that almost nobody else exposes, which matter more on TikTok than on any other platform because sound is how things spread there. Second rather than first only because the team is small, a fact its own FAQ addresses directly with the question "Are you a one-man company?". Fast shipping and a thin bench are the same property from two angles.

3. EnsembleData

TikTok is the platform EnsembleData leads with, and its unit table is unusually legible: one unit buys full user information, or ten posts from a user, or thirty comments on a post, or twenty posts for a hashtag or keyword. Two units buy full post information or a hundred followers. Plans run from a free 50 units a day up to 50,000 a day, and failed calls caused by its own errors are not charged.

The daily cap is the thing to check against your use case. A dashboard pulling a steady trickle fits it perfectly. A one off backfill of two years of a category does not, because you cannot buy past today's ceiling by paying more today, and the plan you would need for one busy week is the plan you then pay for in the quiet ones. It also states plainly on that same page that it scrapes no private or sensitive data and does so "without using fake accounts", which is the specific, checkable claim that puts it above two products further down.

4. Bright Data

The biggest company in the category, with a TikTok scraper sold per record alongside prebuilt datasets and the largest proxy network any of these products sits on. If your requirement is millions of rows with an invoice your finance team recognises and a compliance team you can talk to, it is first rather than fourth and the rest of this list is not really competing.

It is fourth for a smaller buyer because the self serve price sits inside an enterprise sales motion, and because its compliance language, careful and accurate as it is, describes how Bright Data operates rather than transferring anything to you. If you commission the collection of personal data, you are the controller of it, and no vendor's badge changes that. The same caveat applies to every claim on this page and is worth reading once in our Instagram ranking, where the terms question is unpicked in more detail.

5. TikAPI

The specialist. TikAPI does TikTok and nothing else, describes itself as "the only direct data provider" in the space, and covers surfaces the general platforms treat as optional, including live streams, user sessions and the endpoints that only make sense if you have modelled the app rather than the website. For a product whose whole reason to exist is TikTok, that depth is worth paying for.

The reason it is mid table rather than top three is concentration. It is one company, on one platform, and it is the platform most exposed to a change of ownership, a regulatory turn or a client rewrite. Everything above it in this list would survive TikTok becoming difficult by selling you Instagram data instead. This one would not, and if you build a business on it, neither would you.

6 and 7. Decodo and Oxylabs

Both are proxy businesses with scraping products grown over the top, and the ordering explains the ranking. If you already have a TikTok scraper and your problem is that it gets blocked, these are excellent and belong far higher. If you have no scraper, you are buying a component and doing the assembly yourself, which is the exact job the top four exist to remove.

Oxylabs is the larger of the two and the more openly enterprise: a 177 million proxy counter above the fold and a talk to a data expert button next to the free trial. That is not a criticism, it is an address label. Read it as one and you will know quickly whether it is written to you.

8. Supadata, which is doing a different job well

Supadata turns video into text: transcripts, metadata and structured output from YouTube, TikTok, Instagram and Facebook, with a hundred free requests and no card. It is eighth on a scraper ranking because it is not really a scraper, and it is on the list at all because a surprising share of people who go looking for a TikTok scraper want a transcript and have not realised it yet.

The test is simple. If the question is how many views and when, you want one of the top four. If the question is what did they actually say in the first three seconds, and what is the hook doing, no amount of engagement metadata answers it and a transcript answers it immediately. That second question is usually the more useful one, and it is the one behind what a small business should post about.

9. LamaTok

LamaTok is fast, cheap and unusually straight about what it is. Its own structured data calls it an "Unofficial TikTok Private API", the price starts at $0.60 per 1,000 requests, and it says openly that proxies, account rotation, retries and request signing are handled on its side. Its answer to how it differs from the Research API is that TikTok's requires academic affiliation and it does not. There is no pretence here, which is genuinely more useful than a compliance page.

It is last because of what is in its own feature list next to the reading endpoints: DM automation and content upload. A product that can post and send direct messages is not fetching public pages, it is operating accounts, and the free trial being granted after "registration and Telegram verification" rather than a card is consistent with that. If you only ever call the read endpoints you are in roughly the same place as anything else here. The distance between the read endpoints and the write ones is one line of your code, and there is nothing structural stopping the wrong line getting written.

What TikTok data is actually good for

TikTok rewards format more than any other platform, which changes what the data is worth. On Instagram, scraping a competitor mostly tells you what they talk about. On TikTok it tells you what shape works, and shape is copyable in a way that subject matter is not.

• Sound. Which audio a category is using this fortnight is the single most actionable field on the platform, and it is why the sound endpoints matter more than they look.
• Hook structure. First three seconds, first line of on screen text, whether the payoff is at the end or the start. This needs transcripts, not metrics.
• Posting cadence against performance, which is the one competitor number worth having and is discussed for every platform in how often a small business should post.
• Comment mining, which is the most underrated of the four. The questions under a competitor's video are a content calendar somebody else wrote for you.

Notice how little of that needs volume. All four are answerable with a few hundred rows, which is a free tier at three of the vendors above and one afternoon. The subscription only becomes the right shape when the data feeds something that runs continuously.

Before you buy, check whether this is a once

There are good reasons to be here. Building a creator shortlist you will actually contact, sizing a niche before committing a year to it, or feeding a product that genuinely needs social data. For all of those, something in the top four is the right purchase and the free tiers will tell you which within an afternoon.

There is also the reason people do not say out loud, which is that posting has stalled and research feels like progress. Four hundred rows of somebody else's view counts is a satisfying afternoon and it puts nothing on your account on Thursday. If that is nearer the truth, the useful next step is how to use AI for social media marketing or, if the honest problem is that whole weeks pass with nothing published, the category that addresses that is compared in the best AI social media tools.

The same three questions land differently on the other two platforms. Meta has rewritten its terms to close the loophole the courts opened, and YouTube's official API cannot tell you which of its videos is a Short. Those are the Instagram scrapers and the YouTube Shorts scrapers.