Context interview
Tell the platform your question, aims, hypotheses and methods. It warns you when the question is too broad before you start reading.
For people who research and need to publish
You read, think and write. BWL Research handles the rest: it finds the right literature, checks every reference, keeps the page of every citation, runs the statistics and sets up your draft in the journal's format.
Projects/EEG motor imagery
Decoding left vs. right motor imagery from EEG for brain–computer interfaces
WritingAccepted references
23
38 total · 37 verified
Citations
15
Hypotheses
3
Drafts
2
1 pending proposal
Which EEG activity patterns (mu/beta desynchronisation over C3/C4) let us decode whether a person imagines moving the left or the right hand, and which pipeline is most robust across sessions?
Your day to day
The repetitive chores of research happen inside the workflow, and your time goes back to reading, thinking and writing.
Without BWL Research: Searching database after database and copying everything by hand
With BWL Research: One search from several angles across the open OpenAlex catalogue, with a reason for every suggestion
Without BWL Research: Finding out late that a reference doesn't exist, or was made up by an AI
With BWL Research: Every reference is checked in OpenAlex, Crossref, Semantic Scholar, arXiv and PubMed; an unverified one stays out of the paper
Without BWL Research: Reopening the PDF to find the page of a quote
With BWL Research: Every citation keeps its type, page and passage
Without BWL Research: Scattered spreadsheets and statistics nobody can reproduce
With BWL Research: Statistics run by registered code, with tables ready for the paper
Without BWL Research: Fighting the journal template at the last minute
With BWL Research: The draft starts in the venue's format and exports to LaTeX
Without BWL Research: Explaining to the journal how AI was used
With BWL Research: An AI-usage report ready for your submission
How it works
A short interview records your question, hypotheses and methods, and warns you if the question is too broad.
Every suggestion arrives with its reason and its reference checked. You triage into accept, maybe or reject.
Keep citations with their page, note ideas and hypotheses, and connect everything in the argument graph that becomes your outline.
The draft opens in the venue's format, goes through consistency and English review, and exports to LaTeX or Markdown.
A look inside
References being checked, the argument as a graph and citations with their pages, side by side in the same project. Every claim leads back to its source.
Mu rhythm (de)synchronization and EEG single-trial classification of different motor imagery tasks
Addresses hypothesis H2 · uses your declared method (CSP)
EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces
Reference for the baseline of assessment A2
Cross-session motor imagery decoding with transformers
Not found in any open database · stays out of the paper
Features
Literature, citations, ideas, data, experiments and writing in one place. Every claim in your paper leads back to its sources, and every result leads back to the data and code that produced it.
Tell the platform your question, aims, hypotheses and methods. It warns you when the question is too broad before you start reading.
Searches start from the topic, the method, your hypotheses and the seminal work, using the domain's keywords and search archetypes.
Every suggestion says why it is there, such as “addresses hypothesis H2” or “uses the method you declared”. The ranking learns from your decisions.
OpenAlex, Crossref, Semantic Scholar, arXiv and PubMed. Anything that cannot be confirmed is flagged as unverified and stays out of the paper.
Follow the references and citations of accepted papers, and see on the coverage map which areas are still thinly read.
Bring BibTeX, RIS or CSV from the reference manager you already use, triage into accept, maybe and reject, and export as RIS or BibTeX.
Every citation needs a type and a page. Quotes and paraphrases are kept apart from the moment you read, and a correction creates a new version instead of erasing the old one.
Hypotheses, arguments, counter-arguments, questions and gaps, each linked to the citations that support or challenge it.
Proposed, under investigation, supported, refuted or reformulated. The status follows the evidence and your experiment results.
Record the hypothesis under test, the design, the metrics and the data of each experiment, and import results from CSV or spreadsheets.
Tests, confidence intervals, effect sizes and multiple-comparison corrections, computed in an isolated, reproducible sandbox.
An assistant helps you pick the right test, and checks warn you about a missing baseline or unreported variance.
Produced in the publisher's standard, always traceable to the run, the data and the code that generated them.
Register the hypothesis and the analysis plan before the results, with a timestamp and hash that prove what was decided first.
Source, version, licence and sensitivity of every dataset, with a datasheet and lineage, and the exact commit behind each experiment, verified on GitHub or GitLab.
An agent grounded on your lab's knowledge base that cites its sources, is versioned and is evaluated like any other experiment.
Thesis, hypotheses, arguments, evidence and results linked by relations such as “supports” and “contradicts”, laid out across sections in an outline ready to write.
Compare versions of the argument and spot claims without evidence, unused evidence and contributions without a result.
Each section opens next to the citations, ideas and results it needs. NeurIPS ships ready; IEEE, ACM, APA and ABNT are available as lab skills.
Claims without citations, contradictions, unreferenced figures, and an abstract that promises more than the paper delivers.
Suggestions that keep your voice, accepted or declined one by one.
Log reviewer reports and prepare your response, the camera-ready version and the submission package.
Each project knows its phase, suggests the next step and warns you, without blocking, when something important is missing.
One project, several papers: each publication has its own slice of the graph, its venue and its deadlines, and the plan changes as the research does.
Members per project, comments on the draft and versions you can compare.
TOTP with recovery codes, progressive login lockout and session revocation.
Each lab only sees its own data. Sensitive datasets stay in the storage your lab defines; the platform keeps metadata only.
Every piece of content is versioned and audited: who changed what, and when.
LaTeX and Markdown for the paper, BibTeX and RIS for the bibliography, and the AI-usage report for the publisher.
The whole interface speaks Portuguese and English, and everyone picks their own language.
Set up once by the lab's coordination and inherited by every project in the lab.
Versioned guidelines, research lines, funded projects and your lab's skills catalogue.
Scientific base, target publishers with their format rules, and statistical validation criteria that every project inherits.
Domains with inclusion criteria, keywords and synonyms, search archetypes, macro projects and a knowledge base for each domain.
A library shared across projects, from which each researcher brings references into their own work.
Quality indicators per project, such as verified references and claims with evidence, and a live map of how the lab uses the platform.
Results, applications and agents can become candidates for a product, a licence or a spin-off.
AI with authorship
The platform uses AI to search, organise, verify and review, never to take over authorship. Every suggestion arrives as a proposal you accept or decline item by item, and every call is logged.
Suggested ideas, graph links, edits and corrections arrive as proposals. Nothing in your work changes until you accept it.
Paper sections open empty. The AI only drafts text when you explicitly ask for it, with a notice every session.
Every AI call is logged and can be exported as an AI-usage report to declare to your publisher.
Who it's for
From project to publication: verified literature, organised citations and an argument you can see before you write.
Research platform for graduate students
Several papers per project, each with its own slice of the graph, its data and its code, in each venue's format.
Research platform for researchers
Follow each project's phase, comment on drafts, and see where every claim and every number comes from.
Research platform for advisors
Shared guidelines, domains and standards across every project, and knowledge that stays with the group when people move on.
Research platform for labs
Governance, academic integrity and an audit trail for graduate programs and institutes.
Research platform for institutions
Results, applications and agents with provenance, ready to become a product, a licence or a spin-off.
Research-to-product platform
Pricing
The free plan lets you try the platform on a real project. Paid plans add members, projects and daily AI calls.
Free
Free
Try the platform on a real project.
Start free$19/mo
$15.83/mo
$190 billed yearly · save 17%
For researchers taking their work from question to submission.
Start 14-day trial$79/mo
$65.83/mo
$790 billed yearly · save 17%
For research groups running many projects and papers.
Start 14-day trialCustom
Custom
For universities, institutes and graduate programs.
Contact salesPrices in US dollars (USD). Change plans or cancel anytime: your lab keeps all its data.
FAQ
Yes. When you sign up you create your own workspace, which the platform calls a lab. You can invite your advisor and colleagues whenever you like, and the free plan is enough to get started.
A research platform that follows your project from the first question to publication: literature, citations, ideas, data, experiments and writing, all connected and traceable.
No. Paper sections open empty and the AI only drafts text when you explicitly ask for it. Suggested ideas, links, edits and corrections arrive as proposals you accept or decline item by item.
Every reference is checked against open databases: OpenAlex, Crossref, Semantic Scholar, arXiv and PubMed. A reference that cannot be confirmed is flagged as unverified and stays out of the paper.
Yes, as a file: export your library as BibTeX, RIS or CSV and import it into a project or into the lab library. Accepted references can also be exported as RIS or BibTeX.
By registered code running in an isolated, reproducible sandbox, never by a language model. Every table and figure leads back to the run, the data and the code that produced it, and you can pre-register the hypothesis and analysis plan with a timestamp and hash.
Drafts open in the NeurIPS format, and IEEE, ACM, APA and ABNT can be installed as lab skills. Papers export to LaTeX or Markdown, and the bibliography to BibTeX or RIS.
Every AI call is logged together with your acceptance or rejection. That log can be exported as an AI-usage report to accompany your submission.
Each lab is isolated from the others, sign-in has two-step verification, and every piece of content is versioned and audited. Datasets marked as sensitive stay in the storage your lab defines; the platform keeps metadata only.
Yes. Invite members with lab roles, add people to each project, and get comments from co-authors and your advisor on the draft.
Yes, in Portuguese and English. Everyone picks their own language.
There is a free plan to try the platform on a real project, and paid plans for researchers, labs and institutions. Current prices are in the pricing section.
Yes. Cancellation takes effect at the end of the paid period and your lab returns to the free plan without losing any data. Above a plan's limits you only cannot add more of that resource.
Create your account and start with the project you already have. The free plan needs no credit card.
Start free