We use cookies for analytics to improve your experience. You can accept or decline non-essential cookies.

    CreatorsTikTokDebunking

    How to Fact-Check a TikTok Before You Stitch It (and Make Correction Stitches That Land)

    14 min readBy BSmeter

    The short version

    Fact-check a TikTok before you Stitch or Duet it, so your correction lands and you don't get debunked in your own comments. Paste its link at bsmeter.ai and sign in free with Google or email. You get the main checkable claims with timestamps, a 0-100 credibility score for each, a search for evidence sources on the shaky ones and, in Roast mode, a joke-first take on the claims. Then stitch the exact claim, show one receipt, say what's actually true, and roast the claim, not the person.

    How to fact-check a TikTok before you stitch it

    Check the words, not your memory of them. A TikTok you watched once at 1am isn't one you can quote. Here's the check, step by step:

    1. Copy the link

      In TikTok, tap Share, then Copy link. It has to be a public video. A draft has no link yet, so there's nothing to check.

    2. Paste it at bsmeter.ai

      Sign in free with Google or email, paste the link and pick a mode: Roast for joke-first material, Analyze for a neutral breakdown, or Learn if you want the background explained.

    3. Read the claim list

      BSmeter transcribes the TikTok and pulls out the main checkable claims. Each one is pinned to a timestamp and gets a 0-100 credibility score. It also names the move behind the claim, such as Missing Context, Confidence Mismatch, Anonymous Authority or Cherry-Picked, and flags logical fallacies it spots. A claim it didn't flag isn't automatically true.

    4. Open the sources for the low scorers

      For claims that score low, BSmeter runs a search for evidence sources. Click through and read what comes back. Results are AI-generated and can be wrong, and a source is only your receipt if it actually says what you're about to say.

    5. Download the transcript

      Grab the full timestamped transcript as TXT, SRT or VTT. It's free for signed-in users. Now you can find the claim fast and pick the exact stretch to Stitch. The transcript is machine-generated, so replay that moment before you quote it.

    Why check a TikTok before you Stitch or Duet it

    A Stitch puts someone else's clip inside your video. Whatever they said, you're now showing it to your audience. That's the point when you're correcting it. It's a problem when you misheard it, or when you're agreeing with something that turns out to be false.

    • You amplify what you agree with. Stitching "this is so true" onto a bad health tip spreads the tip with your face attached.
    • You mis-correct. Correct the version you remember instead of what they said, and someone in your comments can post the real quote. Now you're the one getting debunked.
    • You pick the wrong target. The loudest line in a TikTok is often the harmless one. The low-scoring claim at 0:40 is the one worth a stitch.
    • You waste your hook. Open with the exact claim and a receipt, not "okay so this is wrong", so viewers know why to keep watching.

    How to make a correction stitch that doesn't get debunked back

    A correction stitch has one job: the viewer sees the claim, then sees why it's wrong, before they scroll. Here's a structure that keeps it tight:

    1. Stitch the exact claim

      Tap Share on their video, then Stitch, and choose the part to use. Stitch is made for a short clip: TikTok's 2020 Stitch announcement said up to five seconds, and its current Stitch help page says you choose the part of the video you want, so check the limit in the app. Use the transcript timestamps to find the seconds where they state the claim. If the whole sentence won't fit, use a Duet, or put the full quote on screen (next step) rather than cutting mid-sentence. Don't trim it so it sounds worse than it is: TikTok's Integrity and Authenticity guidelines count cutting phrases to change meaning as significantly edited content, and viewers can find the full video anyway.

    2. Put the claim on screen, word for word

      Add a text overlay with the quote and where it is, for example "Claim at 0:14". Copy it from the transcript, then replay the moment in their video and fix any word the transcript misheard, so the words you attribute to someone are exactly what they said.

    3. Give the verdict in the first seconds

      One line on what's wrong: "That study was on mice." "That number is from 2009." "He left out the second half of the quote." BSmeter's labels are a good starting point, because Missing Context or Cherry-Picked is often more accurate than just "false".

    4. Show one receipt

      Put one source on screen: who published it, when, and the line that matters. One readable receipt beats five screenshots nobody can read. Name it in your caption too, so people can look it up. Your receipt is the source, not the BSmeter score. The score is an AI-generated pointer to what's worth checking, so don't put it on screen as proof.

    5. Say what's actually true

      Don't stop at "wrong". Give viewers the replacement fact, or say plainly that the evidence is mixed. "Unproven" is a fine verdict when that's what the sources show.

    6. One claim per stitch, then stop

      If the TikTok has four bad claims, stitch the strongest one and save the rest for a follow-up or a Duet. A short, specific correction beats a thorough one nobody finishes.

    Template script: Their clip: the few seconds where they say "[exact claim]". You, with on-screen text "Claim at 0:14: [quote]": "This one's missing context. [Source], [date], says [what it actually says]. So the real answer is [the true version]. Source is in the caption."

    Roast version: Their clip: "[exact claim]". You, with on-screen text "Claim at 0:14": "'Experts say.' Which experts? Anonymous Authority, table for one. [Source], [date], actually says [what it says]. Receipt's in the caption."

    Stitch, Duet or standalone: which format for a TikTok correction or commentary video?

    • Stitch when it's one short claim that fits in a few seconds. Their clip plays, then you. The setup and the punchline stay clean.
    • Duet when the claim runs longer, when you need to respond while they talk (for example a video that makes several claims in a row), or when viewers should see their delivery and your reaction side by side.
    • Standalone when the original comes from a small account or a private person, when replaying the claim is itself the problem, or when the creator has turned Stitch and Duet off. Leave their clip out: state the claim (on screen with its timestamp, or in your own words for a private person), then put the receipt on screen.

    Two catches before you hit record. Not every video allows it: TikTok's help pages say creators can turn Stitch and Duet off for individual posts. Creators can also remove the Stitches and Duets tied to their video, and TikTok's Stitch privacy settings page says doing that also permanently removes their own original post. So put your receipt and source in your own caption too, and if you'd rather your correction didn't depend on their video, post a standalone version.

    Whichever you pick, your part should be most of the video. As of September 2026, TikTok's Integrity and Authenticity guidelines say content that includes unoriginal or reused material without anything new isn't eligible for the For You feed. Adding a real correction is how you bring something new, but For You feed eligibility is still TikTok's call. The same three formats work for any TikTok commentary video; for more angles, see commentary video ideas.

    TikTok's misinformation policy and unverified content: what to know before you stitch (as of September 2026)

    TikTok's current Community Guidelines were released on August 25, 2026 and took effect on September 24, 2026. TikTok updates them regularly, so check the linked pages before you post. These are the rules that matter most for correction and reaction stitches:

    • Harmful misinformation isn't allowed. TikTok doesn't allow misinformation that could cause significant harm. It works with independent fact-checkers and experts, and their assessments feed into moderation decisions (Integrity and Authenticity).
    • Reporting on or condemning misinformation is allowed. The same Integrity and Authenticity page allows documentary or educational content that reports on or condemns misinformation, and TikTok's Enforcement page lists counterspeech and satire among its public interest exceptions. TikTok suggests using captions, stickers or voiceover to make that purpose clear, and says content allowed under an exception can still be kept off the For You feed or get a warning screen or context label. Make your framing obvious from the first second so nobody mistakes your stitch for agreement.
    • Unverified claims can lose reach. Misinformation that may cause moderate harm, unverified information about crises or major civic events, and content temporarily under review by fact-checkers aren't eligible for the For You feed. TikTok may also add warning labels or prompt people to pause before sharing unconfirmed content, and its election rules say it may add warning labels to content its fact-checking partners can't verify (Integrity and Authenticity).
    • Edited clips can need a label. TikTok defines significantly edited content as media that makes it seem like someone did or said something they didn't, for example by cropping or cutting phrases to change meaning, or by rearranging or combining clips. It requires a label (its AIGC label or your own clear caption, sticker or watermark) on AI-generated or significantly edited content that shows realistic-looking people or scenes. Edits that mislead about a matter of public importance aren't allowed even with a label. Small edits like color correction, reframing or cropping don't need disclosure (Integrity and Authenticity). Keep their words intact and in order.
    • Reposts without creative edits lose reach. Reused or unoriginal content posted without creative edits, such as clips showing someone else's watermark or logo, isn't eligible for the For You feed (Integrity and Authenticity). Screen-recording someone's TikTok and re-posting it with a caption is not a correction stitch.
    • Criticize the content, not the body. TikTok's Safety and Civility guidelines don't allow insulting someone's physical attributes, such as weight or facial features. They allow more critical commentary about public figures than private figures (everyone under 18 counts as a private figure) and explicitly allow criticizing someone's content or actions.
    • Strikes and appeals. In a February 2023 newsroom post, TikTok said an account gets a strike when violating content is removed, and that strikes expire after 90 days. The thresholds may have changed since. TikTok's current Enforcement page says you can appeal a removal, a ban or a For You feed restriction, and that For You feed restrictions show up in TikTok's analytics tool.

    How to roast a bad TikTok take without crossing into harassment

    The best correction stitches are funny. The joke just has to land on the claim. "Bro cited a study that doesn't exist" roasts the claim. A joke about his face roasts him, and as of September 2026 TikTok's Safety and Civility guidelines don't allow insulting someone's physical attributes, such as weight or facial features.

    • Aim at the claim, the source and the logic. BSmeter's Roast mode is built this way: it roasts the claims, not people's looks.
    • Go easier on private people. TikTok allows more critical commentary about public figures than private figures, and hostile or profane language aimed at a private figure can be age-restricted and kept off the For You feed (Safety and Civility). For teens and private individuals, skip the stitch. And remember that an analysis of a public link is a public BSmeter page, so don't run a private person's video through Roast mode just for the laughs.
    • Don't send your audience after anyone. TikTok's Safety and Civility guidelines don't allow inciting harassment or coordinated abuse, such as encouraging people to leave abusive comments or falsely report an account. A correction is a reply, not a pile-on.
    • Say "this claim is false", not "this person is a liar". You can show that a claim is wrong. You usually can't show what someone knew or meant.

    What BSmeter can and can't do for TikTok stitches

    • Links it takes: TikTok, Instagram Reels, YouTube (including Shorts), podcasts and public audio-file links. The free plan gives you 3 analyses a day on videos up to 30 minutes; Basic and Premium raise the limits.
    • What you get: the main checkable claims with timestamps, a 0-100 credibility score per claim, the rhetorical move and logical fallacies it spots, and a search for evidence sources on the claims that score low.
    • Transcript: the full timestamped transcript as TXT, SRT or VTT, free when you're signed in. See the TikTok transcript generator.
    • Roast card: a BSmeter-branded image of the result to download, 1080 square or 9:16.
    • On the go: the browser extension shows the analysis on TikTok while you watch, and the Android app takes a TikTok link you share to it.

    What it doesn't do:

    • Check a TikTok draft. It needs a public TikTok link.
    • Clip TikToks to the Doomscroll. Clip is YouTube-only, so for TikToks your output is your notes, the transcript and your own Stitch.
    • Edit video, give you a video file, write your script, or check footage for AI generation or manipulation.
    • Keep a normal analysis private. Analyses are public pages by default.

    Want to check your own correction script before you film it? Check your video privately before you post: only you can see the result, and you can paste your script, an Unlisted YouTube link or a public audio-file link. Making a longer response video? Follow how to debunk a viral video or start from the debunk video tool. More creator workflows: BSmeter for creators.

    Frequently asked questions

    How do I fact-check a TikTok before stitching it?

    Copy the TikTok's link (Share, then Copy link), paste it at bsmeter.ai and sign in free. You get the main checkable claims with timestamps, a 0-100 credibility score for each, and a search for evidence sources on the claims that score low. Open those sources yourself before you stitch, because results are AI-generated and can be wrong.

    Can I check my TikTok video before posting it?

    Not as a TikTok draft. Drafts have no link, and BSmeter needs a public TikTok link. You can check the video you're responding to, and you can check your video privately before you post by pasting your script, an Unlisted YouTube link or a public audio-file link. Only you can see that result.

    Does TikTok fact-check videos?

    As of September 2026, TikTok's Integrity and Authenticity guidelines say it works with independent fact-checkers and experts to assess the accuracy of content and factors their assessments into moderation decisions. Content temporarily under review by fact-checkers isn't eligible for the For You feed, and TikTok may add warning labels or prompt people to pause before sharing unconfirmed content. BSmeter isn't one of TikTok's fact-checkers and isn't connected to TikTok.

    What does it mean if TikTok flags a video as unverified?

    As of September 2026, TikTok's Integrity and Authenticity guidelines say unverified information about crises or major civic events, and content temporarily under review by fact-checkers, isn't eligible for the For You feed. TikTok may also add warning labels or prompt people to pause before sharing unconfirmed content, and its election rules say it may add warning labels to content its fact-checking partners can't verify. Not being eligible for the For You feed isn't the same as removal: TikTok's For You feed Eligibility Standards say people may still find such videos through search or by going to the creator's account. If a video you want to stitch carries a label like that, check its claims before you amplify them.

    How do you make a good correction stitch?

    Stitch the few seconds where they state the claim, put the full quote on screen word for word (checked against their video, not just the transcript), give your verdict in the first seconds, show one clear source, say what's actually true, and cover one claim per stitch. If the claim runs longer than Stitch lets you borrow, use a Duet. Joke about the claim, not the person.

    Is a correction stitch allowed under TikTok's rules?

    As of September 2026, TikTok's Integrity and Authenticity guidelines allow documentary or educational content that reports on or condemns misinformation. TikTok's Enforcement page says content allowed under a public interest exception can still be kept off the For You feed or get a warning screen or context label. Don't cut their words to change the meaning, which TikTok treats as significantly edited content, and don't insult anyone's looks, which its Safety and Civility guidelines don't allow. Enforcement is TikTok's decision, and BSmeter doesn't check videos against its rules.

    Can BSmeter tell if a TikTok is AI-generated?

    No. BSmeter checks what a video says, not whether the footage or voice is real. Look for TikTok's AI-generated content label or a creator's own AI caption (TikTok said in May 2024 it also auto-labels some AI content made on other platforms by reading C2PA Content Credentials). A missing label doesn't prove a video is real. Check the account that posted it and try to find the original upload.

    Is BSmeter free for TikTok videos?

    Yes, with a free account (Google or email): 3 analyses a day on videos up to 30 minutes, plus free transcript downloads. Basic ($3.99/month) gives 15 a day on videos up to 60 minutes; Premium ($8.99/month) handles videos up to 2 hours.

    Sources

    Check the TikTok before you stitch it

    Paste the link, get the claims with timestamps and scores, grab the transcript, and build your correction on receipts you've actually read.