ThinkTwise is an open source browser extension that detects and highlights the quality of arguments in written, digital texts.

  • Constantino Martin Valmaseda, David Noel, Sergio Sanchez

Following the motto “think twice, think wise” the project ThinkTwise aims to create an open source system of applications that reconvert infoxication (also known as information overload) into genuine and trustworthy knowledge. As part of the UNLOCK Accelerator the team of ThinkTwise will develop a browser extension that detects and highlights the quality of arguments in written, digital texts.

Project documentation



Interim report from September 2021:

Showcase at the UNLOCK Demo Day in October 2021:

Project description

What problem does the project solve?

The information overload that we experience online today makes it difficult both to digest the sheer amounts of knowledge users are confronted with daily, but also to properly assess what they are confronted with. Not only keeping right information from wrong is increasingly difficult; additionally, many users are not confident in evaluating the quality of information provided.

What is the intended purpose of the project?

“Our vision is a world where people have access to relevant, new, potent, and trustworthy information and where trust is created through good information, and not institutions.”

ThinkTwise intends to optimize the acquisition of information in written text online. In line with the motto “think twice, think wise” the tool supports people in accessing information with more confidence and in developing a more critical reading attitude. Consequently, users are also nudged to create their own content in a more mindful way. 

The overarching objective is to support a more informed and source-attentive online community that understands what good-quality information is and how to assess it.

How does your project address the problem?

Following a clear predefined set of criteria the ThinkTwise browser extension will allow text highlighting that pinpoints the errors and strengths of written online texts. It is based on  multiple ML/NLP (Natural Language Processing) techniques that detect levels and quality of argumentation in texts, starting by differentiating claims from arguments and then assessing how sounded these claims are.

Further, crowdsourcing and machine learning algorithms will assist the detection of the most informative and validated points, wrong assumptions and possible disagreements.

Which target groups does the project address?

The ThinkTwise browser extension will be designed for two main target groups:

  • general knowledge seekers online sourcing information based on written text, and
  • academics, scientists and researchers specifically whose work is grounded in written texts.


UNLOCK 2021_ThinkTwise__Project logo