Connected Papers
Builds a similarity graph from citation data rather than from a language model, so nothing in the output is generated and there is nothing to hallucinate.
Builds a similarity graph from citation data rather than from a language model, so nothing in the output is generated and there is nothing to hallucinate.
The categories Connected Papers is listed in, and how strong a fit it is in each.
Web research, deep research, books, academic papers, citations, news, fact-checking.
Connected Papers has a free plan. Free plan at $0: 5 graphs a month, with all features included
Not a language model at all โ it draws a similarity graph from citation data. That is why its output can be checked rather than trusted.
Tools that do comparable work, best fit first, drawn from the catalogue rather than hand-picked. Every one is listed whether or not there is money in it for us.
Searches published papers and tells you how many of them support a claim, which turns a yes-or-no question into a count you can inspect.
View Consensus →Searches actual papers and extracts findings into a table you can check, one row per study, rather than answering in prose.
View Elicit →Scoped to clinical evidence and built with practising clinicians, so the filters match how medical literature is actually appraised.
View EvidenceHunt →Runs a multi-step search across dozens of sources and returns a structured, referenced report. Strong first pass on an unfamiliar field.
View Gemini Deep Research →You upload the sources and it answers only from them, linking every sentence to the passage it came from. The strongest verification pattern here.
View Gemini Notebook →Draws a citation map you can steer, so the shape of a literature becomes visible rather than being a list of titles.
View Litmaps →Where Connected Papers sits in the wider category pages.
Built for this — Web research, deep research, books, academic papers, citations, news, fact-checking.
Yes โ Free plan at $0: 5 graphs a month, with all features included. The cheapest paid tier is Academic, at $10/mo quarterly month to month or $6/mo billed yearly.
Builds a similarity graph from citation data rather than from a language model, so nothing in the output is generated and there is nothing to hallucinate. It is listed under Search, Research & Academics, and this is what it is built for.
It finds neighbours of one seed paper, so a bad starting point gives a confidently wrong map. The free tier is 5 graphs a month with every feature included, so paying buys volume rather than capability, and there is no monthly option โ billing is annual or quarterly.
6 tools do comparable work: Consensus, Elicit, EvidenceHunt, Gemini Deep Research, Gemini Notebook, Litmaps. 6 of them have a free plan.
Checked by hand on 4 September 2026. Prices change often โ confirm on the provider's own page before spending.