Papers AI gives researchers and students an AI-native, full-context home for their projects

Researchers and students now have a new AI-native workspace, with the launch of Digital Science’s Papers AI – where they’ll never have to re-explain their work every time they turn to AI for assistance.

Built by the team behind Overleaf, Papers AI is designed specifically for research workflows. Researchers can write, organize their materials and analyze data, with context-aware AI assisting across a whole project, rather than working in one document at a time. From making sense of a paper someone else has written to producing a document of their own, the assistant stays in context throughout the process.

One workspace for the whole project

Papers AI brings research writing, analysis, management and collaborators together in one project, instead of being scattered across separate tools – supporting Word, Markdown, Typst and LaTeX documents alongside Jupyter notebooks, CSV, and Kanban boards.

Papers AI’s fully integrated AI assistant is designed to read a project’s drafts, datasets and references together, reducing the need to re-upload files or re-explain context each time a researcher returns to their work.

Data is never used to train models, and users have the option to use Papers AI completely locally, ensuring no data ever leaves their hardware.

Every AI-suggested change appears as a reviewable edit, which a researcher can accept, reject, or adjust before it becomes part of the document – nothing enters a project unreviewed. When changes are made in one file, other files are updated to reflect those changes, so the whole project stays in sync, with a clear record of what changed and why.

Key features include:

• Multi-format writing – draft and organize in Word, Markdown, Typst, or LaTeX within one project
• AI with full project context – the assistant reads drafts, datasets and references together, with no re-uploading or re-explaining required
• Reviewable AI edits – AI-suggested changes are shown alongside the original, for the researcher to accept, reject, or edit
• Code alongside writing – run Jupyter notebook cells and work with CSV data files in place and reuse outputs directly as figures and tables
• Offline-first and private by design – projects live on-device unless a researcher chooses to share or sync them
• Model flexibility – an external or local AI model can be connected instead of, or alongside, the built-in assistant

Designed with researchers, not just for them

Feedback from the research community has been critical to the development of Papers AI.

Amye Kenall, Chief Product Officer, Digital Science, said: “Researchers lose enormous amounts of time in the gaps between tools – re-explaining a project to an AI assistant that’s already forgotten it, or ferrying results from one system to another. Papers AI supports the whole research workflow in one place, and it never changes anything quietly: every edit is tracked and every file stays in sync, so researchers stay in control of their own work.”

Juan Castro, Principal AI Scientist, Digital Science, said: “Most AI tools only see the page in front of you. Papers AI’s assistant sees the whole project – the draft, the dataset, the notebook, the references – so it can actually help with the parts of research that take the most time, not just the writing. Trust matters as much as capability here – researchers need to know they’re still the ones making the final call on their own work, and that’s something we’ve built into Papers AI from the start.”

Papers AI is available now at papers.ai, with free and paid options, and sign-up handled directly in the app.