IIITH’s Saral Turns Papers Into Easy Videos

SARAL turns research papers into simple videos: Reports

Update: 2025-07-29 16:31 GMT
International Institute of Information Technology Hyderabad (Image credit: Social media)

HYDERABAD: Can research papers be made more understandable to the everyday Indian? A team from the International Institute of Information Technology Hyderabad (IIIT-H) believes so and has built an AI-powered tool called SARAL to prove it. The platform converts complex academic papers into short video presentations, available in multiple Indian languages.

Developed with guidance from the Anusandhan National Research Foundation (ANRF), SARAL simplifies academic papers into a sequence of editable slides and generates corresponding video summaries using text-to-speech engines. The output, a 3 to 4-minute audio-visual presentation, is designed to be accessible not just to researchers but also to students, educators and the broader public.

The idea, according to project lead Prof. Ponnurangam Kumaraguru (Prof. PK), was to democratise research by repackaging it into formats people engage with more easily. “Science communication shouldn’t stop with publication,” he said. “We’re asking how research can be shared, through posters, podcasts or videos, to make breakthroughs more relatable and impactful,” he explained.

The SARAL workflow begins by uploading a research paper, either through LaTeX files, arXiv URLs or PDFs. The tool then extracts core sections, such as introduction, methods, results and conclusion, into slide form, allowing for manual edits and the addition of visuals. Users can select from 11 Indian languages and male or female voices to generate the final narrated video using Sarvam’s text-to-speech technology.

While the current output is a static video, IIIT-H is working on incorporating animation and other formats, such as research posters. The tool is open-source and hosted at democratiseresearch.in, where the team also plans to develop future features aimed at bridging research and industry.

Prof. Kumaraguru’s previous initiatives include AI workshops for PhD students, research mentorship platforms and fellowships like Anveshan Setu. “AI can assist across the entire research lifecycle, not just in doing science, but in sharing it widely,” he noted. The SARAL team now plans to run hackathons and competitions to bring more contributors on board.


Tags:    

Similar News