YP
YouTuberPilot

Moderation that actually understands your comments.

YouTube Studio's filter is keyword-based and silent. YouTuberPilot reads every comment with an LLM, flags toxicity and spam across languages, and gives you one queue for every channel you own — with a full audit trail and push alerts.

No credit card. Requires a YouTube channel.

How it works

From sign-in to a clean comment section in four steps.

  1. 1

    Sign in with Google

    One click with the Google account that owns your channel. We request only the YouTube scopes we need — read comments and (optionally) moderate on your behalf.

  2. 2

    Connect your channel

    Add your @handle and YouTube Data API key in Settings. We pull recent videos and comments from across your channel into one place.

  3. 3

    AI scans every new comment

    Each comment runs through a keyword pre-filter, then an LLM (Gemma) that detects toxicity, spam, scam links, and harassment in any language. Verdicts: clean, flagged, or blocked.

  4. 4

    You review or auto-act

    Flagged and blocked comments land in your queue with a reason. Approve, delete, or reply with one click — or let auto-moderation tiers act for you. Every action is logged.

Why not just use YouTube Studio?

Studio's built-in filter blocks or warns the commenter at post time. It doesn't help you with what actually lands on your channel. Here's what we add:

Context-aware AI moderation

YouTube's filter is keyword + basic ML. We use an LLM that catches sarcasm, coded language, targeted harassment, scam links and non-English slurs YouTube misses.

Unified multi-channel dashboard

Run 3 channels? Studio makes you flip between each one. We show one queue across every channel you own.

Custom rules per creator

"Flag anything mentioning my competitor." "Block crypto/OF spam." "Auto-hide if >3 emojis + link." YouTube only gives you one global toggle.

Kid-safety & parent reports

On-device AI, auto-moderation tiers, and weekly parent reports for families of young creators. YouTube has nothing for this.

Audit trail + history

Every scan logged with a reason ("Contains personal insults/harassment"). Studio silently hides — you never learn the pattern or appeal.

Proactive polling + notifications

Push alert the moment a flagged comment lands, not when you happen to open Studio.

Bulk actions + learning

Approve or delete in batch. The system learns your tolerance so the queue keeps shrinking.

Works where YouTube fails

Hindi/Telugu/Spanish hate speech, leetspeak ("st00pid"), zero-width character attacks — our LLM catches them.

Who it's for

Honest reality: if you're a mega-creator with a full mod team, Studio + ModBot is probably enough. Our sweet spot is creators without that headcount:

What's inside

Ready to clean up your comments?

Sign in with the Google account that owns your channel. We ask for permission to read comments and (optionally) moderate them on your behalf.