Question and protocol
Freeze prior art, claim delta, competing hypotheses, and stopping rules before results
Methods · Calibration · Accountability
We make visible the work between a good question and a trustworthy conclusion.
Finding many papers is not enough. At each stage we separately check fidelity to the source, contrary evidence, numerical and translational accuracy, and caution in clinical use. The method is still being tested across real research tasks; publication status remains explicit until externally reviewed research can be produced reliably.
What can currently be published: Research protocols, open questions, hypotheses, and exploratory syntheses. No work is labelled peer reviewed or a version of record before external assessment.
Nine stages
Nine stages from question to open assessment
Source universe
Build independent lanes for literature, registries, corrections, editions, and cases
Multi-agent screening/extraction
Separate source-criticism, counterevidence, and clinical-translation roles while disclosing correlated model/provider error
Claim Ledger
Trace source → passage → assertion → claim → hypothesis → published sentence
Design-specific appraisal
Apply risk-of-bias, certainty, source criticism, and qualitative methods by study type
Competing synthesis
Preserve supportive and adverse syntheses, divergent predictions, and failure conditions
Reproducible analysis
Prespecification, simulation, discovery/confirmation separation, independent raw-to-figure rerun
Clinical bridge
State patient-important outcomes, harms, reassessment boundaries, and decisions not yet justified
Open adjudication
Publish role-level judgments, counterarguments, version diffs, corrections, and living surveillance
Form before draft
Choose the form that the question can support
The same material demands different validation and supports different conclusions depending on the question. BRC does not force every inquiry into one scholarly form.
What is the unresolved problem?
Open problem / problem essay
States the decision, competing framings, answerability, and stopping conditions.
What has been studied and missed?
Evidence map / scoping review
Shows reproducible retrieval and selection, distribution of sources, and gaps.
Does treatment cause benefit or harm?
Systematic review / research article
Prespecifies comparators, bias, effects, and harms.
Do existing categories obscure the phenomenon?
Critical synthesis / concept-and-method paper
Derives discriminating observations from conceptual history, measurement failure, counterexamples, and rival models.
What does one trajectory reveal?
Case narrative / N-of-1 study
Preserves chronology, context, alternatives, and limits of transfer.
What does a source and its transmission support?
Source / medical-history study
Separates edition, locator, transcription, translation, variants, and reception.
Before drafting, we record the decision at stake, chosen form, rejected alternatives, claims the form cannot support, and redesign triggers.
Figures · Tables · Web
Decide what a display must do before deciding how many to include
Page count is a constraint, not a target. A display remains only when removing it would materially reduce comprehension or auditability.
Illustrated image
When anatomy, a clinical scene, or material texture cannot be expressed adequately with simple geometry
Retain a high-resolution source and disclose generative assistance. Do not use image generation for numerical charts or causal diagrams.
Vector figure
When a conceptual model, anatomical map, process, time course, or variable relationship must be inspected visually
Draw as editable SVG/PDF and test at final PDF size and in grayscale. Automatic tracing is not a scientific master.
Data chart
When real values, distributions, uncertainty, or model output must be compared
Generate reproducibly from data and code, showing units, denominators, and uncertainty.
Semantic table
When definitions, values, sources, predictions, or failure conditions require exact comparison
Preserve rows and columns in HTML, JATS, and PDF. Do not flatten to an image unless historical appearance is the evidence.
Facsimile
When page layout, marginalia, damage, or typography is itself evidence
Identify edition, locator, repository, and rights, with a separate transcription and description.
Prose
When one or two sentences preserve the relationship without loss
Do not add a decorative summary graphic.
Web research lineage
When readers need to explore branching questions, provenance, versions, or audit detail
Evidence required to interpret the PDF is never hidden only on the web.
Research clusters · Weekly release
Release one small question in depth each week
Broad fields such as hot and cold belong to long-running collections. A weekly cluster begins with a narrower phenomenon—back heat, cold extremities, or burning over objectively cool skin—specifying location, felt quality, time and provoking context, and the decision that better knowledge might change. Historical cases, contemporary trajectories, patient language, physiology, medical history, measurement, and counter-hypotheses are connected only when the question earns them.
Symptom-centred work starts with patients’ natural language before diagnostic labels. Location, depth, boundary, timing, triggers, relief, and functional effects are preserved with provenance, while public posts, search queries, and clinical speech remain distinct source lanes. This language atlas generates questions and cognitive-interview items; it does not estimate prevalence or treatment effect.
Weekly describes the public release rhythm, not a seven-day research deadline. Each cluster normally spends two to six weeks or longer in retrieval, close reading, adversarial review, and writing before release. Four weeks are fixed into a monthly archival issue; an empty week is preferable to lowering the publication threshold.
Research tools
Tools widen the search; researchers remain responsible for judgment
Open-copy locators
Unpaywall, OpenAIRE, institutional repositories, and PMC locate lawfully public full text while preserving URL, version, licence, and acquisition hash.
Verified-corpus interrogation
The PaperQA2 adapter admits only artifacts whose acquisition hash matches. Answers and citations remain candidates until passage verification; the runtime is not yet installed or operated.
Expanding the question space
We absorb multi-perspective questioning and concept-map patterns from STORM/Co-STORM without treating generated reports as scholarly sources. Vane, formerly Perplexica, remains only a discovery-interface candidate.
Publication thresholds
A strong average cannot excuse a consequential error
No fabricated source is acceptable, and every central claim must have a verifiable source passage. Numbers and classical-text locations must be reproducible, Korean and English must not diverge in clinically meaningful ways, and material external findings must be resolved or stated as limitations. These thresholds apply to the weakest performance by language, design, and evidence direction—not just to an average.
Calibration cadence
Calibration is continuous
- Every run checks sources, passages, numbers, duplication, retractions, and privacy.
- Monthly tests check whether model or prompt changes reduced accuracy on a fixed validation set.
- Quarterly review includes an external audit and random rechecking of published claims.
- Annual review covers null results, corrections, harm signals, and editorial independence.
A new model is never adopted merely because it writes more fluently. It must demonstrate non-inferiority on fixed benchmarks and improvement in important failure slices.