[1] Garfield, E. Citation indexes for science. Science 122, 108 (1955).

[2] Price, D. J. de S. Networks of scientific papers. Science 149, 510 (1965).

[3] Swanson, D. R. Fish oil, Raynaud's syndrome, and undiscovered public knowledge. Perspect. Biol. Med. 30, 7 (1986).0.91

[4] Torrie, G. M. & Valleau, J. P. Nonphysical sampling distributions in Monte Carlo free-energy estimation: umbrella sampling. J. Comput. Phys. 23, 187 (1977).

[5] Sugita, Y. & Okamoto, Y. Replica-exchange molecular dynamics method for protein folding. Chem. Phys. Lett. 314, 141 (1999).

[6] Laio, A. & Parrinello, M. Escaping free-energy minima. Proc. Natl. Acad. Sci. USA 99, 12562 (2002).

[7] Bolhuis, P. G., Chandler, D., Dellago, C. & Geissler, P. L. Transition path sampling. Annu. Rev. Phys. Chem. 53, 291 (2002).

[8] Mikolov, T., Chen, K., Corrado, G. & Dean, J. Efficient estimation of word representations in vector space. arXiv:1301.3781 (2013).0.84

[9] Rezende, D. J. & Mohamed, S. Variational inference with normalizing flows. Proc. ICML (2015).

[10] Chen, R. T. Q., Rubanova, Y., Bettencourt, J. & Duvenaud, D. Neural ordinary differential equations. NeurIPS (2018).

[11] Noé, F., Olsson, S., Köhler, J. & Wu, H. Boltzmann generators. Science 365, eaaw1147 (2019).

[12] Reimers, N. & Gurevych, I. Sentence-BERT: sentence embeddings using Siamese BERT-networks. Proc. EMNLP (2019).

[13] Karpukhin, V. et al. Dense passage retrieval for open-domain question answering. Proc. EMNLP (2020).0.88

[14] Ho, J., Jain, A. & Abbeel, P. Denoising diffusion probabilistic models. NeurIPS (2020).

[15] Lewis, P. et al. Retrieval-augmented generation for knowledge-intensive NLP tasks. NeurIPS (2020).

[16] Song, Y. et al. Score-based generative modeling through stochastic differential equations. Proc. ICLR (2021).

[17] Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583 (2021).

[18] Albergo, M. S. & Vanden-Eijnden, E. Building normalizing flows with stochastic interpolants. Proc. ICLR (2023).0.79

[19] Lipman, Y., Chen, R. T. Q., Ben-Hamu, H., Nickel, M. & Le, M. Flow matching for generative modeling. Proc. ICLR (2023).

[20] Vaswani, A. et al. Attention is all you need. NeurIPS (2017).

[21] Kuhn, T. S. The Structure of Scientific Revolutions (Univ. Chicago Press, 1962).

[22] Fortunato, S. et al. Science of science. Science 359, eaao0185 (2018).

Literature search for researchers

Eratosthenes: AI literature search and research-gap finder for scientists, ranking papers by meaning and mapping where a field is still open

Finds the closest work to your idea,
and the gaps still open around it.

Named after the chief librarian of the Library of Alexandria, Eratosthenes turns a research question into many optimized academic queries, reads and scores every paper by meaning, and maps the field so you can see where your next contribution fits.

See it in action

Ranked papers Topic maps Research-Gap Graphs PI finder Citation export

Ask. Rank. Map.

One sentence in; a ranked reading list and a map of the field out.

"Flow matching for rare event sampling"

Ranked publications

Stochastic interpolants for rare-event transition paths

Albergo et al.2024142 cites0.91

Flow-matching samplers for metastable molecular systems

Noé Lab202388 cites0.87

Importance sampling via continuous normalizing flows

Chen & Duvenaud2022311 cites0.79

Topic map: clusters and gaps

Generative samplers Rare-event theory MD simulation

How it works

Describe your idea

Type a research question or hypothesis in a sentence, and set the Query Entropy to widen or narrow the net.

Expand and search

The AI expands it into optimized academic queries with synonyms and broader terms, then harvests OpenAlex and arXiv.

Rank by meaning

Every paper is scored against your intent. The AI Match score ranks by conceptual fit, not shared keywords.

Map the gaps

Explore the topic map and the Research-Gap Graph, and find the labs and investigators behind the work.

Not keyword search. A research strategy.

AI query expansion

Your topic becomes a set of optimized academic queries with synonyms and hypernyms. Tune the Query Entropy to widen or narrow the net.

AI relevance match

Each paper is read and scored against your intent, so the closest work rises to the top even when it uses different words.

Topic maps

An interactive map groups the literature into themed branches, from the well-studied core out to the under-explored frontier.

Impact

Research-Gap Graphs

Well-studied topics form deep basins and the barriers between them are the open gaps. Each paper gets an Impact score for how much of a gap it closes.

Author and PI finder

See the principal investigators behind the papers: recent output, lab focus, and fit with your research direction.

Coming soon

Research-plan generator

The most feasible route to close a gap you have found, planned to keep trial and error to a minimum.

Frequently asked questions

Where do the papers come from?

Eratosthenes searches OpenAlex, the open index of scholarly works, and arXiv. Journal articles, conference papers and preprints are all included.

How is this different from a keyword search?

A keyword search returns papers that share your words. Eratosthenes scores how closely the meaning of each paper matches your question, so a relevant paper that uses different terms still ranks near the top.

What is the difference between quick search and deep search?

Quick search ranks up to 60 of the most relevant papers for your query. Deep search reads up to 200, which matters for broad questions and cross-disciplinary topics.

What is a Research-Gap Graph?

It maps the papers in your search as a landscape: topics that are well studied form deep basins, and the barriers between them are the open gaps. Each paper gets an Impact score for how much of a gap it closes.

Is Eratosthenes free?

Yes. Quick search and topic maps are unlimited and free forever, with 3 deep searches and 3 Research-Gap Graphs a month. Pro, Max and Team plans raise those limits; see the pricing page.

Are my searches saved?

Yes. Every search is saved to your account, so you can reopen any past session from the sidebar and export its citations.

Do Eratosthenes and Benjamin share a subscription?

No. You sign in to both with the same SciCoagent account, but each product has its own plan, so you only pay for what you use.

Find the prior work. Spot the gap.

Free to start. Describe your idea and see what Eratosthenes finds.

Eratosthenes: AI literature search and research-gap finder for scientists, ranking papers by meaning and mapping where a field is still open EratosthenesEratosthenes