Information overload is a constant reality. Therefore, tools that facilitate the management and understanding of large volumes of text become essential. EzSum is one such tool, created with the goal of helping students and professionals summarize long documents efficiently.
The idea comes from a personal need. The creator, Miguel Fernández Vivaracho, a university student, was overwhelmed by the number of essays and works he had to summarize. Using ChatGPT for this task, he realized the limitations and costs associated with this tool. “It wouldn’t be bad to have an AI that did the same thing that ChatGPT was doing for me, of course without so many restrictions, not even a clubbing of almost €290 a year that not all students can afford,” says the creator.
Additionally, the summaries generated by GPT did not always meet his expectations, requiring him to be extremely specific in his directions. “In most cases, the summaries that GPT generated for me did not meet my expectations,” he adds. This frustration, combined with a growing interest in programming, led to the development of EzSum. “EzSum is designed by students for students, and that’s why I try to plug all the holes in ChatGPT that I thought of peeking through.”
We spoke with Miguel Fernández Vivaracho, creator of the Ezsum tool, to learn how it works in first person
Technology Behind EzSum
EzSum is not simply a black box that magically generates summaries. The technology behind this tool is based on advanced OpenAI models, but with a specific focus on the task of summarizing long documents. “The key is a process called supervised learning and, in more advanced stages, reinforcement learning with human feedback (RLHF).”
To understand how AI is trained, the creator compares the process with the evolution of the first beings that went from water to land millions of years ago. “It was through trial, error and the identification of patterns, the compilation of tens of millions of years of data, that it managed to adapt to this new medium,” says Fernández. Similarly, AI models start with gigantic databases containing all types of texts, and as they process this information, they develop greater understanding of language.
Additionally, EzSum uses OCR (optical character recognition) technology to process PDFs containing images of scanned text, something many AI tools cannot do. “This ensures that even documents in image format can be summarized easily,” highlights its author.
Differentiation in the Market
EzSum differs from other summary services for several reasons. First, it was created by a student for students, which ensures that the tool is designed to solve problems specific to this group. “EzSum offers flexible and economical packages, including a free plan and single payment options ranging from €1.49 to €9.99,” details Fernández Vivaracho.
In addition, it can handle large documents and offers more than 14 types of specialized summaries, adapting to the user’s specific needs. “It doesn’t matter if it is a novel, a scientific article, a technical report or a marketing strategy, the system adapts to the document and the specific needs of the user,” he adds.
Another difference is the global reach of the platform. EzSum summarizes in 15 languages, such as Chinese, German, French, Arabic and Japanese. “This, combined with our custom-designed prompts, ensures summaries that are not only linguistically correct, but culturally and contextually adapted.”
Document Types and Accuracy
From the beginning, one of the main objectives was to create a tool that was not limited to certain formats or disciplines, but that offered a range as diverse as the needs of those who use it. EzSum can process and summarize virtually any type of document, from academic essays, books and scientific articles, to more specialized documents such as legal contracts, technical manuals, business plans or financial statements.
“It doesn’t matter if it is a novel, a scientific article, a technical report or a marketing strategy, the system adapts to the document and the specific needs of the user,” Miguel Fernández Vivaracho, creator of the Ezsum tool
“To ensure the accuracy and relevance of the summaries, we use carefully designed prompts so that the AI identifies and prioritizes the most important parts of the document,” summarizes the founder of the tool. This guarantees that each summary is not only an abbreviated version of the text, but a synthesis adapted to the characteristics of the document in question.
Primary Users and Access to the Service
Although the tool was initially focused on students, the platform attracts a much broader spectrum of users, including professionals from various fields. “EzSum.me is completely web-based, meaning you don’t need to download anything or install any apps,” says Fernández. This facilitates access to the service, allowing users to obtain summaries quickly and efficiently.
Students find in EzSum an ally to convert their textbooks, class notes, academic articles and even entire theses into something digestible. For their part, professionals use EzSum to analyze contracts or sentences, deal with technical manuals, or process reports and proposals efficiently.
Future Plans
The author of EzSum has big plans for the future of the platform. Among the most notable ideas is the possibility of making summaries of YouTube videos and the incorporation of an integrated PDF editor. “Imagine you’re watching a lecture, a master class, or even a technical tutorial, and instead of taking notes by hand, EzSum can generate a detailed summary of the video content,” he notes.
Another idea is to add an integrated PDF editor within the same platform. Additionally, work is underway to implement advanced fine-tuning, which means further fine-tuning the AI model with specific datasets to improve its accuracy in specific areas, such as medical, legal or financial documents. “The idea is that EzSum is not only versatile, but also unbeatable in specific niches.”
As well as incorporating the ability to summarize audios. “Think about recordings of meetings, interviews or classes: EzSum could transcribe the content and provide a key summary, eliminating the need to review hours of recording,” highlights Fernández. Additionally, the summaries generated by EzSum could be converted to audio format, which would be perfect for those who prefer to listen to a summary while driving in the car, exercising, or simply need a more accessible option.
Finally, the creator plans to further expand the number of document types and summary styles available. “From medical reports to documents for human resources, I want EzSum to be the definitive tool to transform any type of information into something useful and digestible,” he concludes.
