AI-powered and evidence-based research with Consensus

Discover Consensus – the AI-powered research tool that analyses science-based answers from over 200 million research articles and presents them in an understandable way.

Consensus originates from the startup of the same name and uses large language models (LLMs), based partly on OpenAI technology, to analyse scientific papers and present the key findings in a clear and concise manner. It is designed for researchers and students seeking reliable answers to complex questions – directly from the scientific literature.

Consensus Deep Search: Literature Review in Minutes

The new Deep Search feature offers a structured, multi-step research strategy: it identifies foundational literature, highlights opposing viewpoints, and reveals adjacent research areas. Within minutes, a comprehensive AI-generated report is created, including key studies, citations, visualizations, and research gaps.

Consensus GPT: Your Personal Research Assistant

Consensus GPT allows users to ask natural language questions and receive scientifically grounded summaries – complete with citations. This enables efficient validation of ideas, hypothesis testing, and exploration of new research fields.

How to use Consensus

To access Consensus, you must register for a personal, non-transferable account using your ETHZ-email address. Consensus is initially  available to you on a trial basis. Due to the dynamic environment surrounding AI applications, future adjustments to our resources are possible.

Introductory webinar on Consensus

?Consensus 101? is offered as an introductory webinar primarily aimed at faculty members and library staff. It provides an overview of Consensus and how it works, shares best practices and offers an opportunity for questions. The following dates are available:
In English, please register using the respective link.

For more information about Consensus, visit the Consensus information page by the ETH Library.

Although the goal of Consenus is to provide summaries and generative AI functions based on trusted scientific literature, it cannot be ruled out that results may be generated that could be considered false, misleading, biasedor even offensive. Users should not rely solely on the results of generative AI but should conduct independent research and critically evaluate the AI-generated output.

For more information on AI in teaching and learning, visit the corresponding overview page.

The ETH Library is here to support you. Discover our range of courses and workshops

Note on the translation

This text has been translated for your convenience using a machine translation tool. Although reasonable efforts have been made to provide an accurate translation, it may not be perfect. If in doubt, please refer to the German version.

Should you come upon significant translation mistakes, please send a short message to so that we can correct them. Thank you very much.

Always up to date

Would you like to always receive the most important internal information and news from ETH Zurich? Then subscribe to the "internal news" newsletter and visit Staffnet, the information portal for ETH employees.

Similar topics

For researchers

JavaScript has been disabled in your browser