Research intelligence

From local reading to a coherent field.

Every paper you save is a local observation. SHEAF assembles them into the structure of your field: a five-minute daily triage, a live knowledge graph, and literature reviews built from what you actually read.

Get early access → See how it works
Live rendering from SHEAF's graph engine
50,000+ papers mapped nightly 7 quality dimensions per paper arXiv · bioRxiv · PubMed · CrossRef · OpenAlex · Semantic Scholar
Three acts

Daily. Connected. Coherent.

SHEAF moves with your research, from morning triage to structured synthesis. Nothing below is a mockup of an idea: the cards, the graph, and the reviews are the product.

SHEAF. 3 of 5 today
HIGH RELEVANCE
Pass
Save

● Real card, served by a live SHEAF queue this week

Daily.
Five minutes. Five papers. Chosen for you.

Each morning SHEAF ranks the night's new papers against what you've saved before, plus a reserved slot for a strong paper you haven't seen yet, so missing a day never means missing the paper. The act of saving is the act of teaching.

  • Every card explains why it surfaced
  • Summaries written by SHEAF, never pasted abstracts
  • Add any paper by arXiv ID or DOI

● A real collection graph: four saved papers, their shared concepts, and two related papers SHEAF discovered (dashed)

Connected.
Your reading becomes a map.

Saved papers connect through semantic similarity, shared concepts, and author networks: the links a citation graph never shows. Related work you haven't saved appears as ghosts at the edges: discovery, bounded and labelled.

  • Force-directed graph per collection
  • Concept view: methods, datasets, architectures
  • Related papers surfaced from 50k+ mapped nightly
Multi-omics integration: a working review
Literature review · generated from a collection · every claim linked to a source
1 · Problem statement
2 · Early approaches
3 · The foundation-model turn

Illustration. Reviews are generated from your own collections

Coherent.
Start with a map, not a blank page.

One tap turns a collection into a structured literature review, foundations to frontier, with every claim linked back to the paper it came from. Reading paths order any topic by dependency, so you always know what to read first.

  • Hierarchical outline, direct source links
  • Dependency-ordered reading paths
  • Markdown export to your editor
Under the hood

Built like an instrument,
not a feed.

↻

Never lose a paper to timing

Great papers you missed keep competing for weeks. Two queue slots are reserved for the best of your backlog, every day.

⌘

Collections are lenses

Each project gets its own triage feed, its own graph, its own centroid. A paper you add deliberately anchors the whole collection.

∿

Learns what quality means to you

Seven quality dimensions, reweighted from your saves and passes. No settings to tune. Reading is the interface.

⚿

Institutional access built in

Paywalled paper? One tap routes through your university or NHS library. SHEAF never touches your credentials.

⇪

Bring your library with you

Import BibTeX or RIS from Zotero, Mendeley, or EndNote and your graph starts populated, not empty.

§

Original text stays with its authors

Every summary is SHEAF's own writing with a link to the source. We never republish publisher abstracts. Your data is never sold or used for ads.

Membership

Same instrument.
Pay for volume.

Every plan is the full SHEAF: daily triage, the knowledge graph, literature reviews. Plans differ in how much, never in what. Start free; no card required.

Monthly Annual Save 17%
Free
£0
Forever

Find your daily rhythm: 5 papers a day and one literature review a month. Enough to feel what SHEAF does.

Get started →
Basic
£12
per month

The full daily workflow: 25 papers a day, the live knowledge graph, and 10 literature reviews a month.

Pro
£25
per month

Built for heavy use: 50 papers a day, 50 reviews a month, and 20 dependency-ordered reading paths.

Enterprise. For labs and departments: everything in Pro at unlimited volume, plus shared collections, admin controls, SSO, and API access. From £60/seat (5-seat minimum).
Contact sales →
Questions

Asked, answered.

Where do the papers come from?
SHEAF maps new work nightly from arXiv, bioRxiv, PubMed, CrossRef, OpenAlex, and Semantic Scholar. Preprints and journal-published work sit side by side. You can also add any paper directly by arXiv ID or DOI.
How does it learn what I care about?
Entirely from behaviour. Every save and pass updates your taste model; papers you add deliberately count extra; each collection develops its own centre of gravity. There are no sliders or settings to maintain. After around thirty saves the difference is unmistakable.
What happens to a good paper I miss?
It keeps competing. SHEAF ranks a rolling thirty-day window rather than just "what arrived today," and reserves daily queue slots for the strongest papers you haven't seen. Skipping a week doesn't cost you the paper that mattered.
Do you republish paper abstracts?
No. Every summary shown in SHEAF is our own generated text, clearly labelled, with a link to the original on arXiv, bioRxiv, or the publisher. Authors' and publishers' text stays theirs.
What do you do with my data?
Your reading history personalises your own recommendations. That's its job. We never sell data or run ads. Where usage informs product improvement it's aggregated or de-identified, and you can export or erase everything under GDPR. Payments run entirely on Stripe; card details never touch our servers.
Is there an iOS app?
Yes. Daily triage is designed for the phone (currently in TestFlight beta), while the knowledge graph and review builder live in the desktop app. One account, both surfaces.
Can I cancel anytime?
Yes. Self-serve from the billing portal, effective at the end of the period. The free tier is free forever, no card required.
Early access

Reading is the interface.
Start reading better.

We're onboarding researchers in batches. Leave your name and email and we'll send a personal invite when a spot opens.

✓ You're on the list. We'll be in touch soon.

By joining you agree to our Privacy Policy.