Citerra · on writingVol. 01

The all-in-one environment for academic writing

Literature search, a source library, a PDF reader, citations, and cited answers. LaTeX or Word-style, one workspace.

Why do the bands flatten near 1.1 degrees?

3 steps

  • Checked which sources can be cited
  • Searched your library for "band flattening at 1.1 degrees"
  • Read this source

At the first magic angle the interlayer tunnelling and the kinetic scale are comparable, and the Fermi velocity of the lowest bands vanishes.1

1. Moire bands in twisted double-layer graphene

And which of those two measures the angle directly?

GPT 5.6 Luna High80%
FIG. 1Answered from this document's library
§ 1

searches · 600M+ papers, every major database

import · PDF, BibTeX, RIS, DOI, Zotero, Mendeley

reading aids · sections, skim marks, terms, bibliography

the sources

Bring in your sources.

Search PubMed, arXiv, IEEE Xplore, JSTOR and the rest of the literature without leaving the document. One click saves the result, its open-access PDF and its DOI, checked against Crossref.

Or import what you already keep, as a PDF, a BibTeX file, an RIS export, a DOI, or a Zotero or Mendeley collection.

Open one and Citerra marks its sections, its key sentences, the terms it defines, and the works it cites.

A claim belongs to the paper it came from, so the works a source cites are one click from your own library.

This document11 sources
TitleAuthorsYearStatus
Unconventional superconductivity in magic-angle graphene superlatticesCao, Y. et al.2018
Moire bands in twisted double-layer grapheneBistritzer, R. et al.2011
Correlated insulator behaviour at half-filling in magic-angle graphene superlatticesCao, Y. et al.2018
Graphene bilayer with a twist: electronic structureLopes dos Santos, J. M. B. et al.2007
The marvels of moire materialsAndrei, E. Y. et al.2021
Tuning superconductivity in twisted bilayer grapheneYankowitz, M. et al.2019
Superconductors, orbital magnets and correlated states in magic-angle bilayer grapheneLu, X. et al.2019
Emergent ferromagnetism near three-quarters filling in twisted bilayer grapheneSharpe, A. et al.2019
Maximized electron interactions at the magic angle in twisted bilayer grapheneKerelsky, A. et al.2019
Intrinsic quantized anomalous Hall effect in a moire heterostructureSerlin, M. et al.2020
Strange Metal in Magic-Angle Graphene with near Planckian DissipationCao, Y. et al.2020
FIG. 2This document's library
§ 2

scope · this document's library, and nowhere else

review · every change accepted or rejected

the assistant

Write, and let it cite.

Citerra reads the manuscript, searches your library, and searches the literature when you allow it. Then it drafts.

Every claim links to the passage it came from, in the paper it came from. When it cannot ground a point, it says so.

Everything it writes arrives as a change you accept or reject, in your spelling and your citation command.

paper.tex1 change to review
50\end{equation}51%52and the Fermi velocity of the lowest bands vanishes at53a sequence of values of $\alpha$, the first of which is
$\alpha \approx 0.605$.54$\alpha \approx 0.605$~\cite{bistritzer2011moire}.
+1 wordAcceptReject
5556The same twist fixes the size of the moir\'e cell.57With $a$ the graphene lattice constant,58%59\begin{equation}60  L_s = \frac{a}{2 \sin(\theta / 2)},61  \label{eq:period}62\end{equation}
FIG. 3One change, pending
style

+ 37 more

Cao, Y., Fatemi, V., Fang, S., Watanabe, K., Taniguchi, T., Kaxiras, E., & Jarillo-Herrero, P. (2018). Unconventional superconductivity in magic-angle graphene superlattices. Nature, 556(7699), 43–50. https://doi.org/10.1038/nature26160

Cao, Yuan, et al. “Unconventional Superconductivity in Magic-Angle Graphene Superlattices.” Nature, vol. 556, no. 7699, Mar. 2018, pp. 43–50. Crossref, https://doi.org/10.1038/nature26160.

Cao, Yuan, Valla Fatemi, Shiang Fang, et al. 2018. “Unconventional Superconductivity in Magic-Angle Graphene Superlattices.” Nature 556 (7699): 43–50. https://doi.org/10.1038/nature26160.

[1]Y. Cao et al., “Unconventional superconductivity in magic-angle graphene superlattices,” Nature, vol. 556, no. 7699, pp. 43–50, Mar. 2018, doi: 10.1038/nature26160.

1.Cao Y, Fatemi V, Fang S, Watanabe K, Taniguchi T, Kaxiras E, et al. Unconventional superconductivity in magic-angle graphene superlattices. Nature. 2018 Mar 5;556(7699):43–50. doi:10.1038/nature26160

1.Cao, Y. et al. Unconventional superconductivity in magic-angle graphene superlattices. Nature 556, 43–50 (2018).

one record · 10.1038/nature26160

§ 3

reads for · coverage, structure, argument, clarity, evidence

findings · file and line

credit · the original behind the passage, offered beside it

the review

Read it before your reviewer.

Ask for feedback on the whole paper or on one section. Citerra reads all of it either way, and reports what holds before what does not.

It reads for five things: coverage, structure, argument, clarity and evidence. Every statement carrying no citation is checked against your library, and counted.

A finding is a place. Each one carries the file and the line it was found at. Hand the passage to the assistant when you want the fix.

Select a claim and the library answers with the sentence that supports it. When that sentence is citing someone, the original is offered beside it, so the credit lands there.

Feedback
Whole paperRun feedback

Reviewed 3 min agoClear

The flat-band condition is derived before either measurement is used, and the two measurements are the right ones for it. Two statements stand on nothing in the library, and one result is credited to the wrong paper of a pair.

What works

  • The continuum model is stated in full before anything is concluded from it paper.tex:42-54.
  • Every symbol is defined where it first appears paper.tex:42-45.

Coverage1

The half-filling result is credited to the paper beside the one that reports it.

Half filling is credited to the other paper of the pair

Correlated insulator behaviour at half-filling in magic-angle graphene superlattices reports that result, and the note cites Cao et al., Nature 2018 for it, which is the superconductivity paper paper.tex:71-75.

ElaborateFixDismiss

Structure

Sound. Both equations arrive before the sentence that uses them.

Argument

Nothing is concluded that the two cited results do not carry.

FIG. 4One review, as it arrives
§ 4

engines · pdfLaTeX, LaTeX, XeLaTeX, LuaLaTeX

together · carets, comments, tracked changes, history

export · LaTeX, PDF, Word

the editors

Write LaTeX, or Word-style pages.

Write LaTeX with four engines and a TeX Live version per document, compiled on Citerra's own servers. Templates for IEEE, ACM, Springer LNCS, Elsevier and a thesis.

Or start a Word-style document. Same library, same assistant, same citations.

Co-authors write in either at the same time. You see their carets, their comments and their tracked changes, and every earlier version is kept.

When a LaTeX compile fails, Citerra reads the log back in plain words, with a jump to the line and a fix to accept.

paper.texROAFediting now
45The competition is governed byone dimensionless ratio,46%47\begin{equation}48  \alpha = \frac{w}{\hbar v_F  k_\theta},49  \label{eq:alpha}50\end{equation}51%52and the Fermi velocity of thelowest bands vanishes at asequence of values of53$\alpha$, the first of whichis $\alpha \approx 0.605$ and54$\theta \approx 1.1^\circ$~\cite{bistritzer2011moire}.
Handover studyEMJPediting now
Section

3 Method

Twelve handovers were observed on two wards, and each was followed by a short interview with the nurse who gave it. Transcripts were coded by hand twice

“Half of what matters never gets written down. You say it while you walk.”

nurse, ward B

The first pass stayed open. The second kept only the categories both readers agreed on.

Two of the twelve were observed on the same shift, and those two accounts do not agree.

FIG. 5The two writing surfaces

LaTeXpaper.texWord-styleHandover study

§ 5

connect · Zotero, Mendeley, Drive, OneDrive, Dropbox, Notion

sync · GitHub, or any Git remote

sign in · Google, GitHub, ORCID

beyond the tab

Wherever you are reading.

A browser extension saves the paper you are reading to a document's library, with its PDF.

Add-ins for Word and Google Docs bring your citations and the assistant to the writing you do elsewhere.

Connect Zotero, Mendeley, Google Drive, OneDrive, Dropbox and Notion. Citerra is an MCP server, so your own agents reach the same library under the same scope.

Or keep the paper in Git. A repository syncs both ways, so a terminal and the editor land on the same commit.

This documentOne library
libraryZoteroMendeleyGoogle DriveOneDriveODDropboxNotionNbrowser extensionWWordGoogle Docsyour own agents

scope · one document, and the organization it belongs to

FIG. 6Where the library reaches
§ 6 · before you submit

Citerra reads
before it writes.

Citerra drafts from your sources, never from nothing. Everything it writes arrives as a change you accept or reject, and it shows what it read to get there.

  • GDPR
  • EU data residency
  • No training on data
  • Encrypted in transit
  • Encrypted at rest
  • SAML SSO
  • SCIM provisioning
  • Custom roles
  • Audit log

Integrations

Where it connects
  • Zotero
  • Mendeley
  • Google Drive
  • OneDrive
  • Dropbox
  • Notion
  • Word
  • Google Docs
  • OpenAlex
  • Crossref
  • Semantic Scholar
  • arXiv
  • Europe PMC
  • DOI
  • ORCID
  • GitHub
  • Model Context Protocol
§ 7 · pricing

Choose your edition.

Every subscription attaches to an organization. A solo writer, a lab and a faculty use the same Citerra, and the edition opens more of it.

Priced per seat. A subscription belongs to an organization, which adds seats and gets one invoice.

vol. i
Free
Start writing.

  • 10 AI actions a day, on the model Citerra picks
  • Both editors, with compile and PDF preview
  • 100 sources per paper, and no limit on papers
  • Literature search across every major database
  • Journal and conference templates
  • The browser extension
  • One seat

Compare all features

vol. iii
Team
Write with co-authors.
from 3 seats

  • Everything in Pro, per seat
  • 1,200 AI actions a month per seat, pooled
  • No limit on co-authors on a paper
  • A library shared across the group
  • Organization-wide projects and admin
  • Priority support

Compare all features

vol. iv
Enterprise
For an institution.
from 20 seats, priced on request

  • Everything in Team
  • SSO and SCIM
  • Custom roles and an audit log
  • Retention controls
  • Your own model endpoint
  • Institutional billing and a DPA

Compare all features

§ 8 · faqs

Frequently asked questions.

01When does it open, and what happens if I join?
No date is promised. It opens when the writing loop is solid enough to hand over. Joining sends one email to confirm and one when early access opens, and the address goes nowhere else.
02Do I need to know LaTeX?
No. Start a Word-style document instead: your library, your citations, and the formatting your department asks for. If you do write LaTeX, the templates for IEEE, ACM, Springer LNCS and Elsevier hand you a preamble that already compiles.
03Where does my unpublished manuscript live?
In the EU, on a deployment pinned to Ireland. It is never used to train a model. Export any paper as .tex and .bib at any time, or delete the organization with a 30-day grace window.
04Does using Citerra conflict with my journal's AI policy?
That is the journal's call, and disclosure stays yours. Citerra labels and records every AI-authored change, and nothing reaches the manuscript until you accept it.
05Can I bring my Overleaf project across?
Yes. Upload a zip and the .tex, .bib, .cls and figure files land with the folders intact, on the TeX Live version you pick.
06Does it work with my reference manager?
Zotero and Mendeley connect directly, as do Google Drive, OneDrive, Dropbox and Notion. Everything else arrives as an export: EndNote, Papers and Citavi all write BibTeX or RIS, and Citerra imports both. Every paper also keeps a real references.bib in its file tree, so an existing bibliography works unchanged.
07How is this different from ChatGPT, or Overleaf plus a chatbot?
An open-domain model invents a citation to fill a gap. Citerra retrieves from this paper's library only, and a key it cannot resolve there is stripped before the text reaches you. That is how retrieval is built, not a line in a prompt.
08Can I write with co-authors?
Yes. Everyone in the organization edits the same document at the same time. You see their carets, their comments and their tracked changes, and the history keeps every earlier version.

Ask something else

closing remarks

Start with a source.

Open a document, add a reference, and see what changes.

One email to confirm, one when early access opens.