Know what was said around every ad.
Every play of your overlay is recorded. The speech around it is transcribed, translated into English and checked for swearing, then rated for risk per stream and per channel. In any language the stream ran in.
No mut hei, jos tarvitte VPN-ää, niin tosta huutomerkki surfshark, koodilla turtsa neljä kuukautta kaupan päälle.
Mun tekis mieleen antaa tost sulle sakkoa.
Well hey, if you need a VPN, then from that exclamation mark Surfshark, with the code turtsa, four months extra.
I feel like giving you a fine for that.
- Overlay displayedCorrectly
- Brand mentionedYes
From the clip to a rating, play by play.
- 01
Every ad play is recorded
A short clip of the stream is saved around each play of your overlay, on every channel.
- 02
The speech is transcribed
What the streamer says in the clip is turned into text, with timestamps.
- 03
Translated into English
Norwegian, Swedish, Finnish or anything else: you read every stream in English, next to the original.
- 04
Swearing is flagged
Curse words are detected in the original language and marked in the transcript, so nothing hides behind a translation.
- 05
Each play gets a risk rating
The content around the ad is rated no, low, medium or high risk.
- 06
Rolled up per channel
Ratings add up per channel while the campaign runs, so a channel that drifts can be dropped mid-flight.
Most clips come back clean.
Each clip gets a brand safety score and a category. Here a Norwegian Samsung clip is rated No Risk, with the brand mention in the streamer's words highlighted, so you see the safe plays and the flagged ones side by side.
Jeg har byttet til S25 nå, kameraet er helt sinnssykt.
Linken ligger i chatten.
I have switched to the S25 now, the camera is insane.
The link is in the chat.
- Overlay displayedCorrectly
- Brand mentionedYes
Spot the channel that drifts. Drop it mid-flight.
Every rating rolls up per channel while the campaign runs. Most channels sit in no or low risk. When one does not, you see it early, read the transcripts behind it, and take it out of the campaign.
It rates what surrounds your ad, not the streamer as a person.
Example UI with example channels.
Questions about brand safety
What exactly is checked?
The speech in the clip recorded around each ad play. It is transcribed, translated into English and checked for swearing, and the content around the ad is rated no, low, medium or high risk.
Which languages does it handle?
The transcript keeps the original language and adds an English translation next to it, so a Finnish or Swedish stream reads the same way as a Norwegian one. Swearing is detected in the original language.
Does it rate the streamer?
No. It rates what surrounds your ad, play by play. That is the moment your brand is on screen, and the moment that matters to you.
What happens if a channel rates badly?
Ratings roll up per channel while the campaign runs. A channel that drifts toward medium or high risk can be dropped mid-flight, and you can see why in the transcript.
Where do we see it?
In your dashboard during the campaign and in the end report, next to delivery, clicks and chat tracking.
Is the chat checked too?
Chat tracking is a separate view: every brand mention in chat, translated and scored for tone. Brand safety reads what is said on stream around the ad.
See the transcripts from a real campaign.
In a 30 minute demo we open a live dashboard and walk through the brand safety view.