Keeping kids safe on YouTube

How to Stop Your Kid Finding Bad Videos on YouTube

Supervising every minute of YouTube is not realistic, and any guide that pretends otherwise is not written for working parents. This one is. Below is every native option that genuinely helps — explained honestly, gaps included — and the one structural change that stops the problem rather than managing it.

By Watchly Player Staff

Published

The short answer

Turn on Restricted Mode for open YouTube, use YouTube Kids for younger children with 'Approve content' mode enabled, and set up a supervised account through Family Link for older kids. All three are worth doing. None of them removes YouTube's recommendation engine, which is the actual route bad content travels. The structural fix is a curated allow-list like Watchly, where your child can only reach videos you explicitly approved — eliminating the feed that leads them somewhere you would not have chosen.

Free trial • No card required • Cancel anytime

Designed for parent’s sanity and child’s safety Zero ads No algorithm Nothing to install Cancel anytime… but you won’t want to

Every native option — what it covers and where it stops

Restricted Mode is the most commonly recommended setting and the fastest to turn on. On a browser, click your profile picture, scroll to the bottom of the menu, and toggle it on. In the YouTube mobile app, go to profile → Settings → General → Restricted Mode. It applies per device, per browser session, so you need to set it on every screen your child uses. It hides content that has been flagged as mature by YouTube's automated systems or by human review — and it genuinely cuts out a lot. The honest gap: new videos can be publicly visible for hours before classification, and borderline content is frequently never flagged at all.

YouTube Kids is a separate app designed for children up to about twelve. Create a parent account, add a child profile, and choose one of three content levels: Preschool, Younger, or Older. The most effective setting is "Approve content yourself," which turns the app into a parent-controlled library. Only channels and videos you hand-approve are visible to your child. The real limitation is the approval workflow: you approve content one item at a time inside the app, there is no dashboard view, and the interface is designed for young children — which means it often feels babyish to eight- and nine-year-olds who already know what regular YouTube is.

Supervised accounts, managed through Family Link, are designed for the gap between YouTube Kids and independent YouTube use — roughly ages eight to thirteen. You create a Google account for your child, link it in Family Link, and pick one of three content settings for YouTube: Explore (roughly 9+), Explore More (roughly 13+), or Most of YouTube. These are blunt age bands. You cannot whitelist specific channels, you cannot block individual videos, and the underlying recommendation engine is still active inside whichever band you choose.

Set all three for the right ages — they genuinely reduce accidental exposure. But understanding what each one cannot do is what lets you make the right call rather than assuming you are more protected than you are.

The scenarios native controls keep failing

Gaming videos are the most common friction point. A child who loves Minecraft or Roblox will quickly find that the most popular gaming creators on YouTube are adults making content for adults — reaction shots, language, and humour that are not appropriate for an eight-year-old but are not explicitly flagged as mature content either. Restricted Mode does not catch this because the videos are not classified as restricted. Supervised accounts do not catch it because "Explore" is an age band, not a channel filter. The only way to handle gaming videos safely is to specify which creators your child is allowed to watch.

Prank channels are a similar story. A large category of popular prank content involves humiliation, staged arguments, fear reactions, or physical stunts. It is not violent or sexual in the way that triggers automated flagging, but it is not content most parents would choose for their child. It ranks well, it is recommended heavily because it holds attention, and it sits cleanly inside every native safety setting.

"Family-friendly" content that is not. Many creators title and tag their videos as child-appropriate, sometimes cynically. A video titled "Fun science experiment for kids!" might spend two minutes on a legitimate experiment and then pivot to humour, language, or scenarios aimed at adults. Automated systems detect keywords and metadata; they do not watch the video the way a parent would.

Search autocomplete is a specific hazard that rarely gets discussed. Even with Restricted Mode on, YouTube's autocomplete surfaces popular search queries from the full user base, including adults. A child typing a few letters can see suggestions for searches no child should be making — before any filter has a chance to engage. This is not a loophole that will be patched; it is a product feature that reflects the full search index.

Why a curated library is a different model, not a stricter filter

Every native option described above is a filter-and-block system sitting on top of an open platform. The platform stays open, the filters try to intercept the worst of what it surfaces, and the recommendation engine keeps finding new paths. You end up in a maintenance loop: periodically checking what your child watched, occasionally resetting the algorithm, turning on a setting you heard about, hoping nothing slips through. This is not a criticism of parents who do this — it is the only set of tools the native platform offers.

A curated library is a different architecture. Instead of starting with all of YouTube and blocking the bad parts, you start with nothing and add only what you have approved. The recommendation engine becomes irrelevant because it has no feed to populate. Your child cannot stumble across a prank channel because there is no browse surface where one could appear. Search autocomplete only shows content already in the library. The question changes from "did anything bad get through today?" to "is there anything new I want to add this week?"

This is the model Watchly is built on. There is no "Up Next" chosen by an algorithm, no sidebar of related videos, and no open YouTube search. Your child's home screen is a library of channels and videos you added, or imported from a curated set organised by age and topic. When a video ends, the screen returns to that library. An AI scan runs against every video's transcript and captions, flagging content from approved channels that drifts off-tone — so even a creator you trusted last year gets reviewed when they post something different.

One gap to be clear about: live gaming streams with uncaptioned voice commentary are outside what any transcript-based scan can catch. If a streamer speaks in real time without subtitles, that content cannot be reviewed before playback. For families where live streams are part of the approved library, that is worth knowing. For the much larger category of standard uploads and pre-recorded content, the scan covers it.

Or skip the workarounds

Watchly is the allow-list YouTube never gave you: only the videos you approved, on any device.

Start Free Trial

Free trial • No card required • Cancel anytime

Native YouTube safety settings vs Watchly allow-list

Native YouTube settingsWatchly
Blocks new uploads before classification No — visible until flagged Not needed — only approved content exists
Per-channel approval YouTube Kids "approve" mode only — slow and per-item Add entire channels from the dashboard in seconds
Recommendation engine active Yes — on all three native platforms No — there is no feed to recommend into
Safe search autocomplete Autocomplete draws from full YouTube query index Search only covers the library you approved
Content scanning after approval Filter applied at upload time only — no re-checking AI transcript scan flags off-tone uploads for review
Works for 8–12 age gap (too old for Kids, too young for open YouTube) Blunt age bands only — no per-creator control Tailored library grows with the child at any age

How Watchly closes the gap native settings leave open

Approved content only. No algorithm. No surprises.

Allow-list by default

Everything on YouTube is blocked. The only videos your child can reach are the ones you explicitly added — channels, playlists, or individual videos.

Library-only search

Search runs against your approved library, not all of YouTube. Autocomplete only surfaces content you have already approved.

AI content scan on every video

Every approved video is checked against 21 content patterns — profanity, violence, scary themes, and more — so off-tone uploads surface for your review first.

Import by age and topic

Start with a curated library matched to your child's age and interests, then add or remove channels from the parent dashboard at any time.

Full watch-history dashboard

Every video your child watched is listed by date and duration in the parent dashboard — on your account, where your child cannot clear it.

Enforced daily limits and breaks

Set a daily time limit and intermission breaks per child profile. The session ends at the limit; the library does not keep queuing up content to extend it.

This is what it looks like

Tap any screen to try the real thing.

What Kids Can't See

No Comments. No Ads. No Algorithm.

No comments section. No algorithm. No autoplay rabbit holes. Just safe content that you've approved.

  • Comments completely removed
  • Parent-approved content only
  • No algorithm manipulation
  • Time limits that work
Parental Controls

Limits You Set Once

Give each child a daily watch time, a bedtime window and intermission breaks partway through. The app enforces all three, so the end of a session is not something your kid negotiates with you.

  • Daily watch time per child
  • A bedtime window that closes on its own
  • Intermission breaks partway through
  • Enforced by the app, on every device
See how the controls work

How to set it up in Watchly Player

1

Add your child's profile

Create a profile with a name, emoji avatar, 4-digit PIN, and birth year. Each child gets their own library, time limit, and watch history — completely separate from siblings.

2

Approve the library

Search YouTube from the Watchly dashboard, preview any video before adding it, and import a curated age-and-topic library to fill gaps quickly. Everything your child watches is something you looked at first.

3

Set limits and let them watch

Set the daily time limit, enable intermission breaks if you want them, and hand over the screen. The library is the entire universe your child can reach — no algorithm, no drift, no surprises.

Start Free Trial

Free trial • No card required • Cancel anytime

27%

of YouTube videos watched by kids 8 and under are made for older audiences.

Common Sense Media & Michigan Medicine, 2020

No algorithm

No feed, no recommendation rail, no autoplay. Inside Watchly there’s nowhere else to go — just the library you built.

Zero ads

No pre-roll, no mid-roll, no banners, no sponsored anything. The video you approved plays — and nothing else.

Designed for parent’s sanity and child’s safety

Part of that: we never sell or rent your data, or your children’s. It’s in our privacy policy in capital letters.

Questions parents ask about this

The honest answers we give in our parent support channel.

My child knows how to clear their watch history. Does that make monitoring pointless?
On native YouTube, largely yes — history is easy to clear, and a determined child will figure it out. It is worth being honest about this rather than pretending it is a security control. Watchly's watch-history dashboard is on the parent account, not the child's profile, so your child cannot clear or access it. You see every video watched, regardless of what the child does on their own screen. That is a meaningful difference.
What is the single most dangerous gap in Restricted Mode?
New uploads. When a video is uploaded to YouTube it can be publicly visible for hours or days before automated systems or human reviewers have categorised it. During that window, Restricted Mode cannot hide it because it has not yet been classified. Creators who know this can also exploit it: upload a borderline video, collect views and engagement before it is flagged, then repeat. Neither of these gaps is theoretical — they are well-documented behaviours, and they are the reason Restricted Mode is a useful coarse filter but not a guarantee.
We already have YouTube Kids but my child says it is for babies. What do we do?
This is one of the most common situations parents describe, usually around ages seven to ten. YouTube Kids does feel pitched at younger children, and a child who knows regular YouTube exists will often resist it. The practical middle path is a curated library that carries the specific creators they actually love — gaming channels, science channels, whatever holds their interest — without opening the full platform. A child who watches their favourite Minecraft creator in Watchly rarely complains it feels babyish, because the content itself is not.
Search autocomplete showed my child something awful. How do I prevent that?
YouTube's search autocomplete reflects popular queries across all users, including adults. Even with Restricted Mode on, the autocomplete suggestions themselves are drawn from the broader search index and can surface queries your child should not see. The only way to fully prevent this is to remove open YouTube search from the child's experience entirely. On YouTube Kids in "approve content" mode, search is replaced with a search of your approved content. In Watchly, search works against the library you built — not all of YouTube — so autocomplete only ever shows content you have already approved.
Are there any gaps in Watchly I should know about before switching?
Yes. Watchly scans captions and transcripts against 21 content patterns to flag off-tone uploads from approved channels before your child sees them. The honest gap is live gaming streams with uncaptioned voice chat: if a streamer is speaking in real time without subtitles, no transcript-based tool can catch what is said. This affects a small subset of content — live streams with active voice commentary — and is worth factoring in if your child is into that specific format. Pre-recorded content, standard uploads, and captioned videos are all covered.

Set it up once, stop worrying

Free trial. No card. Your first child profile and a fully approved library are ready in under five minutes.

Start Free Trial

Setup in 2 minutes • Free trial • Cancel anytime