DCFN - Research is the Living Profile of a research corpus — where the field came from, where it's converging, where the gaps are. It doesn't generate connections; it discovers them structurally. It pulls real papers — hundreds of them — from seven sources, builds a concept graph in memory, and runs five cognitive operations that trace how ideas relate, where knowledge is missing, and where independent researchers are converging without knowing it. It separates what's structurally ready to build on from what needs new research. Not a search engine. A research portfolio manager that tells you where to invest effort and where the ground is already solid.
Try It Now Learn MoreEach source is deduplicated and merged. Full text retrieved when legally available.
200M+ papers with citation graphs and reference chains
Primary36M+ biomedical and life sciences articles with MeSH classification
Peer-Reviewed2.4M+ preprints across STEM with full-text access
Preprints250M+ works with concept tagging and institutional data
Open AccessCode repositories, READMEs, and technical documentation from public projects
Code & DocsML model cards, dataset descriptions, and research artifacts from the open-source AI ecosystem
AI / MLUpload your own paper to see where it sits in the landscape
Phase 2Each run retrieves up to ~900 merged articles per domain across the seven sources. Licensed deployments scale beyond this ceiling.
DCFN - Research finds bridges between research areas. Enter 2 or 3 topics — like “formative assessment” or “circadian rhythm” — and DCFN - Research pulls hundreds of papers from seven sources, builds the concept graph, and maps where the fields converge.
Upload your draft — see where your work sits in the landscape
Enter at least 2 research topics
AI research tools accept a query and return a synthesis — that part is commodity. Perplexity summarizes what papers say. Gemini generates plausible connections. DCFN - Research is different in what it does with the corpus: graph-native structural mapping — typed edges, entropy scoring, citation chains, convergence anchors — surfacing the Living Profile of a field: where it came from, where it's converging, where the gaps are. Patterns no amount of reading would reveal. Three layers of depth. The query intake isn't the differentiator. The structural reasoning over the corpus is.
DCFN - Research builds a concept graph from real published research and runs five cognitive operations across it. It traces backward through citation lineage, projects forward to find where knowledge should cascade but doesn't, detects entropy gaps where a concept should exist given the surrounding structure, identifies bridge concepts connecting disparate fields, finds independent researchers converging on the same territory without citing each other, and surfaces a set of novel research directions — each grounded in a specific structural gap, not language patterns.
Every Article ships with a paired Technical Report — a per-run receipt covering what executed, what it produced, and how it knows. Run fingerprint, stage timings, and patent-mapped operation traces, linked directly from the Article.
LAYER 2 — GAP RESEARCH
DCFN - Research ranks the entropy gaps and untested foundations from your Article by structural severity. It formulates targeted queries for the worst gaps, pulls new papers from all seven sources, merges them into the graph, rebuilds, and computes a before/after structural diff.
LAYER 3 — AUTONOMOUS TRAJECTORY
DCFN - Research stops following your query and follows the evidence. Calibration memory, bridge priors, co-occurrence patterns, and convergence signals guide autonomous exploration into territory no single query could reach. DCFN - Research decides what to research next.