Predict maladaptive behavior with AI-powered analysis of written communication

Dechefr [de · che · pher] is an advanced risk assessment solution designed to provide guidance and support when analyzing written communication by identifying warning signs and risk indicators. By combining linguistics, psychology, and computer science, Dechefr constructs a psychological profile that enables the prediction of extreme or violent behavior with high accuracy.

Read more about Dechefr

Fast threat assessment with high accuracy

Speed up threat assessment and reduce workload for threat assessment teams, forensic linguists, psychologists and security experts. Use Dechefr as an assistant for reliable, automatic linguistic and psychological text analysis that saves hours of work and reveals where further action is needed.

Get free demo access

Use any text source

Automatically evaluate text gathered from various sources such as social media, chat rooms, forum posts and e-mails to get a preliminary risk level and an assessment using Dechefr’s key risk indicators.

Connect linguistics with psychology

Dechefr's unique automated threat assessment complements linguistic analysis with a psychological signature, proven to predict the likelihood of violent behavior.

Compare results with known perpetrators

All risk indicators are presented in a standardized format, with the possibility to compare results against the normal population, known perpetrators, and cases in your Dechefr case archive.

Get evidence-based predictions

Dechefr is validated to predict violent behavior with high accuracy. Risk score prediction and individual risk indicators are designed to work with written language, based on a subset of established structured professional judgment instruments such as TRAP-18.

Combine automatic and manual threat assessment

Dechefr leverages a set of indicators that only requires written communication for structured analysis. The indicators can be assessed both using Dechefr and manually. This allows you to use Dechefr for structured results, ready for verification — or as your fast track to a second opinion.

Get free demo access

  • Quickly see what’s going on. Free up time for in-depth analysis with automatic extraction of named entities, frequently used words and word colocation. Text characteristics such as readability score plus average word and sentence length are available at a glance.

  • Pinpoint threats in need of escalated response and use your resources wisely. Dechefr points out elevated risk through an automatically computed risk level. To guide investigation, risk indicators can be explored and evaluated manually.

  • Monitor threat levels for online forums and individuals. Dechefr can help you through large volumes of text and produce a structured threat assessment. Results are presented in a standardized format to speed up comparisons.

  • Gain clues hidden in text to support assessments and reveal more details. In situations when the person of interest is not at hand, Dechefr can use text sources to analyse indicators associated with violent behavior.

  • Get the basic details served. Dechefr reduces time spent on fundamental text analysis while revealing the psychology behind the words. Need to add your own indicators? Get in touch! Dechefr can be tailored to spot the behaviors and characteristics needed to drive your assessments forward.

  • Dechefr is pronounced as de · che · pher. The word dechefr origins from the French word déchiffrer (decipher).