Twincler keeps platforms healthy at scale
Twincler is a fully automated SaaS abuse intelligence system.
It's arguably the fastest technology to identify platform abuse and harmful content at scale.
Our technology runs fully automated across languages, countries and platforms. It does not need human analysts to identify or review flagged content.
Our customers
Content platforms for counter-abuse and identity teams, intelligence desks, trust & safety teams, recommendation & ranking systems
National security agencies for enhanced situational awareness (OSINT)
Available for
X
YouTube
Instagram
TikTok
Mastodon
Look at the threat not the feed
We understand harmful content and abusive behavior on content platforms as data anomalies in the data stream.
Our systems identify these anomalies fully automatically by content-agnostic pattern recognition and send alerts to customers with contextual information in real-time.
Core advantages
Ahead of unknown threats
Twincler detects previously unknown tactics, abuse types and narratives fully automatically faster than industry-leading systems.
Context in real-time
Easier lead processing and decision-making with additional layers of contextual information in real-time such as associated narratives, account attribution, off-platform amplification or past activity and ownership.
Reducing the noise
Reduce the number of leads for your analysts and intelligence desks with actionable high-accuracy leads.
Content-agnostic flagging
We detect harmful content and abuse content-agnostic by pattern recognition and prevent biased decision making.
Evidence beyond content
Receive data evidence of cross-platform activity related to content and accounts in the data as it streams.
Background: Problems with legacy technology
Problem: Content-based systems
Content-based systems are notoriously expensive and slow to build, are prone to label biases and have limited ability to adapt to new abuses or changing policies.
Problem: Virality and reach based detection
Virality and reach approaches ignore nuances of different types of amplification, are blind toward harmful content that is not boosted, and identify harmful content too late.