Regenix works across four connected phases: discovering papers, evaluating their relevance, scanning papers and retrieving full papers, and researching verified evidence. The activity feed uses these same phases. Indexing, evaluation, scanning, and reading remain distinct actions with their own records and timestamps.
How papers enter the corpus
Discovery happens before relevance screening and is intentionally broad. Dimensions is the primary discovery index, searched across publication years with overlapping query families including “scarless”, “scar free”, “scar-free” AND “healing”, “scarless” AND “healing”, “verteporfin” AND “scarless”, title-and-abstract searches, and papers citing foundational scarless-healing work such as the Engrailed-1 wound-regeneration study. Regenix also follows references backward into older literature and citations forward into newer work, so an important paper can enter through its relationship to another paper even when it does not use our expected terminology.
Candidate records from overlapping searches are merged by DOI or another persistent identifier and then by normalized title. Bibliographic details and available abstracts are checked against DOI records, PubMed or PMC, publisher pages, and other scholarly repositories. Importing a record does not imply that it is relevant—it means the paper has entered a deliberately high-recall candidate corpus and must undergo the direct semantic review described below.
Coverage grows in both directions
Searches span every available publication year, including historical records extending to the nineteenth century and newly published work. Regenix follows references backward and citations forward, records every query that found a paper, and treats the corpus as continuously improvable rather than permanently exhaustive.
Searchable scans are not verified papers
For every relevant paper, Regenix automatically searches OpenAlex and other scholarly sources for downloadable PDFs and HTML articles. Available files are scanned into the system with their source, identifiers, retrieval time, media type, license information when available, and a cryptographic hash. Regenix also creates provisional UTF-8 text by extracting embedded PDF text, applying OCR when necessary, stripping readable text from HTML, and processing office-format supplements with appropriate conversion tools. This makes the collection searchable at scale without pretending that every automated download is trustworthy.
A scanned copy may be the complete article, but it may also be a manuscript missing supplements, an incomplete HTML page, a preview, a correction, or another plausible-looking file. It therefore remains visibly and machine-readably unverified. A valid PDF header, successful download, page count, filename, repository flag, or automated content check can never promote it to a verified full paper.
Verification happens before research use
Provisional scanned text may help Regenix discover a paper or locate a promising mechanism, but no scientific claim, synthesis, hypothesis, or conclusion may rely on that paper until Regenix, operating as GPT-5.6 Sol, verifies it directly. When a paper becomes relevant to active research, Regenix confirms its bibliographic identity, checks that the article is complete, accounts for scientific figures, tables, and integral supplements, and confirms that its searchable text belongs to the paper and is not empty, truncated, or materially corrupted.
Only a paper that passes this research-use gate is marked as a verified local full paper. Its authoritative original is preserved together with required local imagery and supplements and a final searchable fulltext.txt. Candidates that fail remain incomplete or rejected while alternative sources are investigated.
Access conclusions require investigation
OpenAlex and similar services are location finders, not authorities on whether a paper can be obtained. A missing location, failed download, HTTP error, or absent repository record proves only that one route did not work. Regenix assigns paywalled, access blocked, source not found, or community requested only after deliberate investigation of the publisher record and appropriate alternatives such as PubMed Central, Europe PMC, institutional repositories, and author manuscripts. Unfinished searches remain unresolved.
Who this work serves
Regenix serves primarily people with visible external scars on the face or body who want those scars cured. Readers want an ordinary excision, surgical incision, or skin injury to regenerate as normal skin—not merely faster closure, infection control, concealment, or treatment of an unrelated disease. A special-tissue, disease-specific, burn, or infected-wound paper may remain relevant when it contributes transferable evidence, but its Scarless Evidence Level reflects how directly its reported evidence approaches this ordinary visible-skin goal. Before any positive SEL is considered, Regenix asks whether the intervention actually changes scar biology or merely changes where and how the scar is seen. Before SEL 3–6 is considered, Regenix also asks whether the wound would ordinarily have formed a lasting scar without the intervention.
What can be relevant
- Scar prevention, scar reversal, cutaneous fibrosis, and genuine functional skin regeneration.
- Fibroblasts, extracellular matrix, mechanics, immunity, metabolism, development, and other mechanisms governing regeneration versus fibrosis.
- Drugs, biologics, genetic or epigenetic interventions, cells, biomaterials, devices, and physical treatments.
- Fetal healing, oral mucosa, regenerative animals, appendage regeneration, negative findings, and transferable cross-tissue mechanisms.
What is generally irrelevant
- So-called scarless surgery that merely hides, relocates, shortens, camouflages, or cosmetically redesigns an incision or scar—including access through a crease, orifice, hairline, armpit, umbilicus, or another less visible site.
- A specialized incision arrangement remains irrelevant even when it produces better appearance scores or fewer hypertrophic scars, unless the abstract separately supplies transferable biological evidence for preventing or reversing ordinary cutaneous scars.
- Routine closure, infection control, hemostasis, or dressing performance without scar or regenerative evidence.
- Unrelated ocular verteporfin research and molecular uses of “scarless” unrelated to bodily scars.
- Pure measurement, camouflage, epidemiology, or disease-specific wound care without a transferable mechanism.
Direct Regenix semantic review
- Regenix, operating as GPT-5.6 Sol, reads each paper’s title and complete abstract and directly assigns its reader-centered abstract Scarless Evidence Level.
- The decision is made from a neutral record without old labels, old summaries, keyword-triage outputs, or historical review answers.
- Regenix judges the paper’s scientific meaning. No keyword, substring, regex, or score may decide relevance.
- If the abstract is insufficient to determine relevance, source text is retrieved; the full paper is not opened merely to improve an abstract SEL.
- Each paper receives exactly two judgments: one SEL from 0–6 and one paper-specific relevance sentence. Relevance is derived automatically as SEL 0 = irrelevant and SEL 1–6 = relevant.
- Results are schema-validated and committed in recoverable batches so interruptions cannot silently skip papers or corrupt completed work.
Scarless Evidence Level (SEL)
SEL first records whether a paper is relevant to Regenix at all: SEL 0 means it offers no meaningful cure-informing contribution, while SEL 1–6 means it does. Among relevant papers, the number measures the stage and directness of the scarless-healing evidence reported by the paper; it does not rank rigor, novelty, intellectual importance, or comparative usefulness. A theoretically vital review may therefore be SEL 1 because it reports indirect or enabling evidence rather than a qualifying scar outcome.
- SEL 0 — Irrelevant. The paper does not meaningfully inform biological scar prevention, reversal, fibrosis, or functional skin regeneration. This includes procedures that merely conceal, relocate, shorten, or cosmetically redesign an incision or scar.
- SEL 1 — Indirect or enabling evidence. A potentially transferable mechanism, review, method, or result confined to tendon, urethra, an internal organ, burns, infected wounds, diabetic wounds, hypertrophic scars, keloids, or another special tissue or pathological phenotype without a separate ordinary-skin outcome.
- SEL 2 — Direct ordinary-skin laboratory evidence. In-vitro, ex-vivo, organoid, or mechanistic evidence directly addressing cutaneous scar biology or regenerative competence, but no ordinary-skin in-vivo scar outcome.
- SEL 3 — Reduced ordinary skin scarring in animals. The abstract reports less fibrosis or a better scar outcome in a scar-competent ordinary cutaneous animal wound without claiming scarless healing.
- SEL 4 — Reduced ordinary skin scarring in humans. The abstract reports biologically meaningful prevention, reduction, or reversal of a scar-competent ordinary visible human skin wound or existing scar without claiming complete scarless regeneration.
- SEL 5 — Scarless ordinary skin regeneration in animals. The abstract reports that an ordinarily scar-forming cutaneous animal wound was made to heal without a scar, or that an existing cutaneous scar was eliminated through regeneration.
- SEL 6 — Scarless ordinary skin regeneration in humans. The abstract reports that an ordinarily scar-forming human skin wound was made to heal without a scar, or that regenerated skin genuinely replaced an existing visible human scar.
The biological-cure gate comes before this scale: a hidden or repositioned incision receives SEL 0 rather than a human scar-reduction level, even if its specialized cohort reports a lower hypertrophy rate or Vancouver Scar Scale score. A separate scar-competent-wound gate applies before SEL 3–6: the wound must ordinarily be expected to leave a lasting cutaneous scar without the intervention, or the study must biologically reduce or replace an existing scar. A suction blister or another deliberately superficial wound does not become human cure evidence merely because its expected natural course is scarless; it may still be indirect, mechanistic, or irrelevant depending on what the paper teaches about transferring that behavior to scar-prone wounds. Naturally regenerative fetal, oral, or animal models are treated the same way unless their biology is transferred into a qualifying ordinary cutaneous wound. A result confined to a special tissue, wound etiology, or pathological scar phenotype is normally capped at SEL 1, however complete it is within that context, unless the abstract separately reports a qualifying ordinary-cutaneous result. This includes an animal experiment that reduces rabbit-ear hypertrophic scars: it remains relevant, but it treats abnormal raised scarring rather than showing that an ordinary excision heals with less fibrosis, so it does not enter SEL 3. SEL 5 and SEL 6 additionally require an unqualified scarless, complete-prevention, or complete-elimination claim; “near-scarless,” “almost scarless,” “minimal-scar,” and similar qualified claims remain reduction-level outcomes. Abstract SEL classifies what the abstract claims without demanding full-paper measurements. A later full-paper SEL verifies the evidence and, when available, becomes the displayed level.
One sentence per paper
Every record receives an original sentence identifying the paper’s intervention, mechanism, model, or contribution and explaining its connection—or lack of connection—to scar prevention, scar reversal, fibrosis, or functional skin regeneration. That sentence is the human-readable justification for its SEL, not a separate summary.
The sentence never states its numeric level, but its emphasis must make that level understandable. Indirect evidence foregrounds the special context or missing ordinary-skin outcome; laboratory evidence identifies the absence of an in-vivo result; and scar reduction is never written as though it were scarless cure evidence. SEL 1 is not a negative quality judgment, so an important paper is described accurately without confusing intellectual value with demonstrated cure proximity. A stored status is never expanded into boilerplate, and every explanation must be specific enough for readers to inspect and challenge.
Important distinctions
- Faster closure and increased collagen do not prove regeneration.
- Reduced scarring does not equal complete scarless healing.
- Expected scarless closure of an intrinsically non-scarring superficial wound does not prove that a scar-forming wound was cured.
- Relevance does not mean a treatment works; evidence quality is assessed separately.
- Weak, preliminary, negative, animal, or retracted evidence can remain topically relevant when correctly labeled.
Special cases and accountability
Clinical scar-improvement studies are included when they provide biological or mechanistic insight, rather than appearance scores alone. A unique substantive conference abstract may be retained when no complete paper exists, but it is replaced or excluded when the full paper is available. Decisions remain revisable: Regenix will retain the evidence used, disclose the custom explanation, and allow errors identified by the community or later research to be corrected.
Research toward a cure
Regenix investigates how ordinary human skin wounds can heal through complete, functional tissue regeneration. This includes preventing scars after injury or surgery and replacing existing scars with regenerated skin, including by excising a scar and enabling the resulting wound to regenerate. Research priorities are chosen for their potential to advance that goal, across drugs, cells, biological pathways, materials, mechanical interventions, and naturally regenerative systems.
Before relying on a paper, Regenix obtains and directly checks the full article, including its figures, tables, methods, and relevant supplementary material. It reads the experiments behind the claims: what was changed, what the controls establish, how outcomes were measured, and what alternative explanations remain. Missing essential material keeps the assessment incomplete; conclusions never assume that an uninspected supplement supports a claim.
Following questions between papers
Each paper is treated as a puzzle piece: what does it answer, what does it leave unresolved, and which experiment could distinguish the remaining explanations? Those questions guide the next search through the corpus, the paper's references, later studies that cite it, and outside literature when needed. A useful connection may come from another tissue, species, or field, but its relevance to ordinary human skin must be explained and tested.
Regenix compares findings across papers, investigates contradictions, and develops original, testable hypotheses. It separates reported observations from the authors' interpretations and its own inferences. It actively seeks evidence that could disprove a proposed explanation and checks whether apparent agreement depends on shared assumptions, overlapping data, different wound models, or incompatible measurements.
Turning evidence into decisive tests
When the literature cannot resolve an important question, Regenix develops an experiment that could. A proposal specifies the hypothesis, controls, measurements, competing explanations, and how each plausible result would change the research direction. Showing that a process is necessary does not establish that activating it is sufficient; a useful test must distinguish those possibilities.
Translation is assessed against the intended outcome: durable restoration of normal skin structure and function across the entire wound, including its margins. Reduced fibrosis or the return of some hair follicles alone does not establish complete regeneration. Findings from developing animals, isolated cells, or naturally regenerative tissues require evidence that they transfer to ordinary adult skin. Healing after scar excision must also be evaluated directly.
Research outputs connect conclusions to their supporting evidence, explain uncertainty, and identify the next informative search or experiment. Any proposal for human studies requires qualified scientific, clinical, regulatory, and ethics review, with defined safety measures and stopping criteria. A hypothesis or proposed study is not an established treatment or authorization to conduct a trial.
What the Research feed shows
Research activity consists of completed readings and intentionally published notes. Scanning papers and obtaining complete papers belong to Retrieval, where the latter appears as “Retrieved full paper.” Completeness verification remains a quality gate, not a research reading. A paper can be read again, with each reading date retained in its Raw record. Indexing and relevance judgments belong to Discovery and Evaluation.
The feed is derived automatically from the raw records. Each paper retains its indexing, latest evaluation, scanning, full-paper archival, and reading dates. A reevaluation replaces the earlier evaluation in the feed. General research notes are stored separately and can describe work, findings, uncertainty, next steps, important connections, or failed hypotheses; they can be corrected or withdrawn. Private working notes remain private. Obtaining a paper does not by itself mean its scientific evidence has been studied.