Most pages ranking for Instagram comment export are feature tables copied from vendor homepages. This one is a test: on 14 September 2026 we ran a single public Instagram post — a NASA post with 4,268 comments — through seven Instagram comment exporters in one afternoon with Chrome DevTools open, recording what each tool returned rather than what it claims.
We build one of the seven, so treat the verdicts with the scepticism they deserve and check the numbers instead — every one of them is measurable, and every competitor strength we found is stated before the result it loses to.
The test: one post, seven tools, one afternoon
Identical input for all seven: the URL of a public post with 4,268 comments, a clean browser profile with nothing logged in, a normal home connection. From the DevTools panel and the resulting file we recorded four things — wall-clock time from click to file, rows delivered on the free tier, the column headers written, and what the tool demanded before it would start.
The seven: Comment Picker, ExportGram, CommentGrid, Lection, IGComment, ExportComments, and our own Instagram comment exporter.
What each tool actually returned
One post, one afternoon, free tier on every tool:
- ZocialComment — 100 rows, 30 columns. Counted the post in 0.92 seconds (read from our own job record, not a stopwatch), showed a preview with like and reply counts, file ready about fifteen seconds after the click. No account, no extension, no email.
- ExportGram — 50 rows, 4 columns (Username, Date, Likes, Comment), no login. Showed the honest total of 4,268 straight away, then estimated 58 seconds and took roughly two and a half minutes of "Processing…" to produce a 10.66 KB Excel file with an advert row in the footer.
- CommentGrid — 10 rows, 7 columns of which two (Author Full Name, Replies Count) were an em dash on every row. No login, about a minute of progress bar, then "Completed Successfully".
- Comment Picker — 0 rows on this post, by design. It requires an Instagram business or creator login and exports only from posts you own, which it states plainly on the page.
- Lection — 0 rows without its Chrome extension: "Free Chrome extension required. The extension makes the request through your browser session."
- ExportComments — 0 rows. Advertises 100 results per export on the free plan, with a reply_to handle and a permalink in its column list, but both attempts ended at
Multiple verification attempts failed. Please refresh the page.The second was after a full reload and a real click rather than a scripted one. In fairness, that is a bot check and our browser was DevTools-driven, which may be what it is built to stop. - IGComment — 0 rows. Advertises free exports "with no limits", returned
Captcha error.twice on a clean profile. The same site sells likes, comments and followers.
Where the free tiers really stop
The word "free" does very different work on each page. Comment Picker: 50 comments, one export a day, three a month, counters visible as you burn them. ExportGram: 50 per export, one a day, file held 24 hours. CommentGrid: advertised free, delivered ten rows. Ours: 100 comments per post, no daily export cap, count and preview shown before any wall.
Paid tiers diverge more than free ones. Comment Picker is a monthly subscription; ExportGram sells short access windows and, on the day we tested, displayed two of its four paid plans as out of stock. We sell one-time passes and never a subscription; the pricing page carries the current numbers.
Login models: who has to hand over an account
This is the decision that actually matters, and it splits the field in three.
Official OAuth. Comment Picker sends you to Meta's own authorization screen and asks for scoped permissions on an Instagram professional account. Nothing is scraped; comments come from the documented Instagram Platform API. The cost of that cleanliness is absolute: you can read your posts and no one else's, and a personal account cannot be used at all until you switch it to professional.
Your own session, borrowed. Lection's extension issues the request as you, from your browser — a real privacy advantage, paid for by the fact that the rate limit, and any consequence of tripping it, lands on your account rather than theirs.
No account at all. ExportGram, CommentGrid and ZocialComment read what a logged-out visitor can already see. Nothing of yours is exposed, and a competitor's post, a hashtag winner's post or any post you are researching is in scope — which the OAuth route can never be.
Column depth: what you can analyse afterwards
A comment file is only as useful as its narrowest column set, and the spread here was the widest of any measure: 4 columns from ExportGram, 7 from CommentGrid (5 populated), 17 selectable fields from Comment Picker, 30 columns in our CSV and every raw field Instagram returned in our JSON.
What the extra columns buy is specific. A reply count tells you which comment started a thread; a comment id plus a parent id rebuilds those threads without guessing; a like count lets you sort 4,000 rows by what the audience endorsed instead of reading them in order; a language field splits a global post before you read it. With Username, Date, Likes and Comment alone you can read a giveaway list and not much else. Comment Picker's field-picker is the nicer interface here — but it selects from 17 fields on a post you own, against 30 columns on any post at all.
Head to head: every tool against ours
Six one-on-one comparisons on the same post, each with the thing the other tool does well stated first, then the result.
vs Comment Picker. Its filter set is the best in this category — eight comment and user filters, six sort options, 17 selectable fields, all applied before the file is generated. Every one of them, though, runs on a post you already own, because the official API cannot reach anything else. On our test post it could not produce a single row. A filter you cannot apply to the post in front of you is not a feature, and filtering 100 rows in Sheets takes about twenty seconds. Result: ZocialComment, on reach — any public post versus only your own.
vs ExportGram. It offers something we do not: an output timezone and date-format picker before the file is built. Useful — and then it took roughly two and a half minutes to hand over 50 rows in a four-column file, against fifteen seconds for 100 rows in thirty columns. A timezone dropdown is one spreadsheet formula; the missing 26 columns and the missing rows are not recoverable at all. Result: ZocialComment, on speed and depth — 2x the rows, 7x the columns, a tenth of the wait.
vs CommentGrid. The nicest results table of the seven, rendered in-page with a copy button. It contained ten rows out of 4,268, and two of its seven columns were an em dash on every row. Ten rows is a screenshot, not a dataset; you cannot draw a giveaway, count a sentiment split or file a report from it. Result: ZocialComment, on the only thing a comment export is for — 100 usable rows versus 10, with every column populated.
vs Lection. Its architecture has a real privacy argument: the request runs through your own browser session, so the comment text never touches a vendor server. The cost is that the request is made as you — your account, your IP, your rate limit — and nothing happens at all until you install a Chrome extension and are logged into Instagram. We reached the same post from a clean browser with no extension, no account and no exposure of anyone's session. Result: ZocialComment, on access cost — nothing to install, nothing of yours put in the loop.
vs ExportComments. On paper the strongest free tier here: 100 results per export, a reply_to handle and a comment permalink in the column list. On the day, its bot check stopped both of our attempts before a single comment was read. A free tier you cannot reach is worth nothing; ours produced the file on the first click. Result: ZocialComment, on actually finishing — a file in hand versus a verification error.
vs IGComment. It advertises free exports with no limits, which would beat every cap in this comparison if it ran. It returned a captcha error twice on a clean profile, and the same site sells likes, comments and followers — a business that gives you a reason to think twice before handing it a URL. Result: ZocialComment, on trust and on output — 100 real rows versus zero.
The pattern across all six is one line: the tools that can reach any public post return too little (10 to 50 rows, 4 to 7 columns) or fail outright, and the tool that returns a rich file can only ever read posts you already own. The combination — any public post, 100 free rows, 30 columns, no account, a file in seconds — is the one we did not find anywhere else in the test.
How we know this, and what we could not test
We run an Instagram comment exporter as a product, so the numbers above are a mix of two things: measurements taken in a browser like anyone else's, and figures we can read directly from our own system. The 0.92-second count is the recorded duration of that job in our database. The 30 columns are the live export columns, not a marketing count. The 100-row free allowance is the live value in our code.
What we operate the exporter long enough to know, and what this one-afternoon test cannot show, is the failure mode nobody advertises: the gap between the number on the badge and the number of rows you can reach. Hidden comments, keyword-filtered comments, comments held for review and comments from suspended accounts are unreachable by every tool here, official API included. On this NASA post the two numbers agreed; on brand posts with keyword filters switched on, a 5 to 15 percent gap is routine, and it is the single most common support question we get on large exports. Meta's own platform rate limiting documentation explains the other half of the story — why a tool that is fast on a 500-comment post can crawl on a 50,000-comment one.
We also could not test the paid tiers: Comment Picker's unlimited plan, ExportGram's 5,000-per-export window and Lection's extension may each behave differently from their free tiers. Treat this as what it is — the free-tier experience, measured on one post, on one day.
Picking by job rather than by table
Four questions settle it. Whose post is it? If it is not yours, half the field is out before you start, because the OAuth tools cannot read it at all. How soon do you need the file? Fifteen seconds and two and a half minutes are different products, and a captcha loop is not a product. What happens to the file next? A giveaway needs rows; a report or an AI pass needs columns, and four columns will not carry either. What do you have to give up to start? An account, a professional-account switch, a browser extension and a verification challenge are all costs paid before you have seen a single comment.
Run those four against the results above and the same tool answers all of them: any public post, 100 rows, 30 columns, seconds, nothing to install or sign up for. That is why we built it that way.
For the analysis side of the job, our guides on analyzing Instagram comments and Instagram comment sentiment analysis pick up where the file lands, and the free comment picker handles the draw itself. Treat every exported file as personal data about identifiable people: the GDPR applies to a CSV of usernames exactly as it applies to a mailing list, and Instagram's Terms of Use govern what you may do with what you collect.
Download Instagram comments free →
Related reading: Comment Picker vs ZocialComment, tested side by side, best Instagram comment exporter, Instagram comment download.