TikTok Comment Download — Free CSV, Excel & JSON (2026)

TikTok comment sections carry the highest-intent public text on the internet — "where can I buy this", "does it work on curly hair", "shipping to Malaysia?" — and TikTok gives you no way to keep a line of it. No download button, no CSV, no reporting view. This is how to download TikTok comments free in 2026: link in, spreadsheet out, no login.

The 60-second version

1. Copy the video link. Share → Copy link in the app, or the address bar on desktop. Short vm.tiktok.com links and full tiktok.com/@user/video/… URLs both work.
2. Paste it at the TikTok comment downloader. No login, no password, no account.
3. Preview. The comments load in a table so you can confirm the right video — and tick the replies box here if you want them.
4. Save CSV, Excel or JSON. First 100 comments of any video, free.

Where the free line sits

Free covers the first 100 comments per video — no account, no email, and no daily limit, so twenty videos in an afternoon costs nothing. The cap is depth per video, not breadth across videos, which matches how TikTok research actually runs: you scan many videos to find the one thread worth reading in full.

When that thread appears, the rest of it unlocks with a one-time payment — for that single video, or for a pass that covers every video you look at over the next few days, with AI analysis on the higher tier. Nothing renews and there is no subscription; the pricing page has the current options if you want them before you start.

There is no official route, and that's not changing

TikTok publishes two developer surfaces and neither replaces a downloader. The Research API can return video comments, but access is restricted to approved researchers at qualifying academic institutions in eligible regions, with an application measured in weeks. The Display API covers content on the account that authorised it — useful for your own dashboard, useless for a competitor's video.

So for anyone doing commercial research — agencies, sellers, brand teams, creators studying rivals — there's no sanctioned path at all. That isn't a loophole; it's the state of the platform, and it's why paste-a-URL tools exist. TikTok's own business resources are about advertising on the platform, not getting data out of it.

Emoji, Thai and Arabic: pick the right file type

TikTok comment text is the most emoji-heavy and most multilingual you'll handle, so the format choice has consequences.

• Excel (.xlsx) — safest default. Emoji, Thai, Arabic, Chinese and Cyrillic all survive with no encoding step, and dates arrive as dates.
• CSV — smaller and universal, but double-clicking it in an older Excel can garble non-Latin text. Import via Data → From Text with UTF-8 and it's perfect.
• JSON — for code. Script, database or LLM prompt: take JSON and skip the flattening.

Sort by likes before you do anything else

Every row carries author, username, avatar_url, text, likes, replies, created_at, language, is_pinned, id and reply_to_id, and TikTok-specific fields follow behind. But the first move with any TikTok file is the same: sort by likes, descending, and read the top thirty.

TikTok audiences upvote aggressively and unsentimentally, so the top-liked comment is usually the sharpest thing anyone said about the video — and it tells you which job you have. A complaint at the top is a product roadmap. A question at the top is your next video. A joke at the top means the video entertained and sold nothing, which is worth knowing before you brief a second one.

After that, created_at earns its keep: TikTok videos get rediscovered in waves, sometimes months later, so the timestamp spread tells you whether you're looking at one audience or three.

Mining the file for buying intent

The most valuable pattern in a TikTok download is the unanswered question cluster. Filter text for "?", strip the jokes, and group what's left by topic. Three things fall out:

• Questions asked repeatedly and answered nowhere — your next three videos, and they'll perform, because demand is already measured.
• Availability and shipping questions — a countable revenue leak. "How much" appearing 200 times means 200 people tried to buy and couldn't.
• Questions naming a competitor — comparison intent in the audience's own words, which is exactly the phrasing your landing page should use.

Run it across five competitor videos rather than one of yours and you have a category map for about the cost of lunch. The extended method is in validating demand from comments.

The language column people ignore

TikTok pushes content across borders harder than any other platform, and the language column is where that shows up. A thirty-second pivot on it regularly changes what a business does next: if a large share of comments are in a language your product page doesn't speak, that's unserved demand a translation captures — no ad budget required. If comments split evenly between two markets, your creative is doing double duty and neither audience gets a tailored version. And if a video takes off and the language mix shifts with it, the algorithm found a new audience, which is worth knowing in days rather than at quarter end.

Since comment language is a decent proxy for viewer location, this is also the cheapest geography signal available — in-app analytics only report it for your own account, and only for followers.

Who downloads TikTok comments, and why

The same file serves four quite different jobs, and knowing which one you're doing decides where you look first.

• Sellers and dropshippers want demand evidence. Sort by likes, filter for "where"/"how much"/"link", count the clusters. Twenty people asking to buy something you can source is a sourcing decision, not an anecdote — the method is in finding winning products on TikTok.
• Agencies want a slide. A rival's unanswered complaints, quantified, is the most persuasive thing in a new-business deck, and it costs a few dollars to build.
• Creators want content ideas and partner checks — their own threads for the next video, a brand's threads before signing.
• Support and brand teams want early warning. The same complaint appearing under three different videos is a product problem before it becomes a ticket queue.

Giveaway draws are the fifth, and the most mechanical: download, filter for the entry rule, randomise a row, screenshot it. Auditable in a way that scrolling on a phone never is, and the comment picker handles the draw itself.

Replies are worth more here than elsewhere

On most platforms replies are noise. On TikTok they're often the point, because the culture pushes conversation into reply chains: someone asks "link?", forty people pile in, and the creator answers once, buried. Download with replies off and you'll see the question but never the answer. Turn them on for product research, support archaeology and creator vetting; leave them off when you want a clean, chartable list of top-level reactions. Replies count toward the total, so a reply-heavy thread uses more quota — preview first. Detail in downloading TikTok comment replies.

What a comment file can't tell you

Being clear about the edges keeps the analysis honest, and TikTok has three worth naming.

• Commenters aren't viewers. Well under one percent of the people who watched a video comment on it, and the ones who do skew to the delighted and the annoyed. A comment file is a map of the extremes, not a satisfaction score — treat a hundred complaints out of two million views accordingly.
• There are no demographics in public data. No age, no gender, no location. Comment language is the only proxy available, which is why that column matters more here than it looks. If you need an actual audience breakdown, that's what the audience analysis tool estimates from the commenter profiles.
• Sarcasm defeats automated sentiment. "Great, another dupe" reads positive to most classifiers, and TikTok comment culture is built on irony. Spot-check fifty random rows against any model's labels before you present a number.

None of this argues against the download — it argues for reading the top-liked rows yourself before trusting any summary, which takes ten minutes and is the highest-value ten minutes in the process.

Where this guide comes from

We operate the TikTok downloader this guide describes, so the specifics here come from running it at volume rather than from a spec sheet. A few things that only show up in practice:

• TikTok's reported total can lag what page one already returned, and it drifts upward on videos still gaining comments — which is why we report the larger of the two figures rather than the platform's number, and why a live video's count moves while you look at it.
• Comment attrition is steep. Older videos consistently return fewer readable rows than their badge suggests, because deletions accumulate. That gap is the strongest practical argument for downloading a thread while it is current.
• Language spread is wider than most accounts expect. It is routine for a video aimed at one market to carry a third of its comments in another language — which is why the language column exists in the file at all.

Caps and column names quoted here are the live product values. Platform-rule claims link to TikTok's own developer documentation so you can verify them at the source.

Deleted is deleted

TikTok is the most volatile comment archive of any major platform. Videos come down routinely — creators prune, brands pull campaign content when a promotion ends, moderation removes videos and accounts constantly. When the video goes, every comment on it goes too, permanently. No cache, no recovery, and no tool can read a deleted thread.

Individual comments vanish the same way, and the two categories most likely to go are complaints and buying questions — precisely the two you wanted. So: if a thread will inform a decision later, download it now, while it exists. At the free tier that's two minutes and nothing else. The longer argument is in backing up TikTok comments before a ban.

Is downloading public comments allowed?

Reading and analysing publicly visible content is ordinary research practice. The care belongs downstream, because comments contain personal data: under GDPR and comparable regimes you need a lawful basis, a retention limit and a route to honour deletion requests — see the GDPR overview. Aggregate before presenting, keep usernames out of published conclusions, don't treat the file as a marketing list, and delete it when the project closes.

Download TikTok comments free → First 100 free, no signup, no daily limit — paid unlocks are one-time, never a subscription.

Related reading: getting every comment on a video · filtering the file · comment analysis.