Sharly AI: an AI research assistant for source-grounded document analysis
Researchers increasingly use AI assistants on their own documents, but the answers rarely show which passage they came from, and journals now require any such use to be disclosed. This note describes Sharly AI, version 2026.09, as a research instrument, so that authors can use, cite and disclose it. The description is drawn from the production system as deployed on 11 September 2026 and from published uses of the tool. Sharly retrieves passages from the documents a user uploads, has a large language model write an answer from them, and shows up to three source excerpts, with page numbers for PDFs, that the reader checks against the original. As of September 2026, eighteen works refer to Sharly, eleven of which used it in their methods. With the verification protocol given here, Sharly is suitable for literature screening, data extraction and content analysis; it is not a substitute for reading the sources.
Authors
- Simone Macario (ORCID: https://orcid.org/0009-0007-5479-3703)
Institutions
- Voxtel (United States) (US)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-09-17
- DOI
- https://doi.org/10.5281/zenodo.22803981
- Primary Topic
- Computational and Text Analysis Methods
- Type
- article
- Field-Weighted Citation Impact
- 0.00