Asta
Status:
watch-list| Maturity:emerging| Last reviewed: 2026-03-04
What it does
Asta is a scholarly research assistant developed by Allen Institute for AI (Ai2) that combines literature understanding with data-driven discovery. It uses 108M+ abstracts and 12M+ full-text papers to find, summarise, and analyse scientific evidence, bridging the gap between paper search and data analysis.
Key capabilities
- Literature search grounded in Ai2’s 108M+ abstract corpus
- Full-text analysis across 12M+ papers
- Scientific evidence synthesis and summarisation
- Data-driven discovery across the research literature
Fit for Sheffield researchers
Primarily suited to STEM and health researchers given the corpus composition. The combination of literature search and data analysis in one tool is distinctive. Worth monitoring as the product matures — currently in early stages.
Licensing and access
Access model and pricing not fully established at time of review. Recommend monitoring for institutional licensing options as the tool matures.
Trust and governance notes
US data residency (Allen Institute for AI). GDPR compliance status unclear. Not recommended for sensitive research data until governance is confirmed.
Observed limitations
- Early-stage product with limited track record
- Corpus weighted toward STEM — less useful for arts and humanities
- Limited community feedback available at time of review
Related tools
semantic-scholar · elicit · julius-ai
Sources and evidence
- Ai2 Asta product page: https://asta.allen.ai/ (accessed 2026-03-04)