8 June 2026Research Intelligence
/ product

Eight features. One coherent workspace.

Each feature ships against the same primitive: the verified-quote stamp. Search a paper, open it, quote it, cite it, monitor it, validate it. The same provenance follows the quote from PubMed to your manuscript.

FEATURE · 01

Verified Quotes.

Every quote is hashed against a page-anchored region in the source PDF before it ever reaches your document. No quote, no claim.

How it works

On retrieval, Docling extracts a page-anchored quote region. The agent's quote is matched verbatim to that region. Mismatches are rejected before the synthesis step. The hash and timestamp are persisted alongside the quote in your Markdown.

hfref-review.md · DAPA-HF · §2
● local · 0 egress
VERIFIED · 4
McMurray 2019
Packer 2020
Zinman 2015
Marso 2016
In DAPA-HF, dapagliflozin reduced cardiovascular death in HFrEF[1]
“the risk of worsening heart failure or death from cardiovascular causes was lower among those who received dapagliflozin”
McMurray et al. 2019 · p.19991
FEATURE · 02

Citation Chaining.

Traverse cited-by and cites two hops out. Rerank with MedCPT. Build a 47-node working graph from a single seed paper.

How it works

The Literature Reviewer issues graph.expand against the PubMed-derived citation network. Each node is rescored with MedCPT-Article against your question; nodes below threshold are dropped. The PRISMA log records every visited node and the inclusion/exclusion decision.

Graph inspector · DAPA-HF
● 47 nodes · 112 edges
GRAPH · 47/112
depth2
cited-by87
cites25
FEATURE · 03

Multimodal PDF understanding.

Docling extracts text, figures, tables, and page-anchored regions in one pass. Quote regions are stored with byte offsets, not just page numbers.

How it works

docling.parse runs locally on the PDF. The output is a structured document tree: text spans, figure regions, table regions, all anchored to (page, paragraph, character) coordinates. Quotes resolve to specific regions; figures embed in your synthesis with their original captions.

Frangoul et al. 2021 · NEJM
● docling · 312ms
Frangoul H et al. NEJM 2021 · p.252Figure 2
Fetal hemoglobin · CTX001 vs baseline
EXTRACTED · DOCLING
page252 ¶6
figurefig2_hb_pancellular.png
tabletbl1_baseline.csv
verbatimverified · 312ms
FEATURE · 04

Hybrid Retrieval.

BM25 first-pass over PubMed and your corpus. Dense rerank with MedCPT-Article. HyDE rewriting for short queries. The retrieval inspector shows the per-stage contribution to every result.

How it works

Three stages, observable. BM25 narrows the candidate pool; dense reranking applies a biomedical cross-encoder; HyDE generates a hypothetical answer to widen short, ambiguous queries. Each stage's contribution is logged to the run record so you can debug a missed paper.

Retrieval inspector · live
● local · medcpt-article
RETRIEVAL · “GLP-1 vs SGLT2 MACE”
BM25
124
DENSE · MedCPT
18
HyDE rewrite
12
RERANK · top-k
8
FEATURE · 05

Local-First Workspace.

Documents live as Markdown on disk. Embeddings, reranking, and optional LLM inference run on your hardware. Egress is opt-in, per call.

How it works

A standard project folder: manuscripts as Markdown, corpus as PDFs, embeddings in Qdrant on disk, retrieval models as gguf binaries. The Control Plane shows you which tools will leave the machine before each call. Diff your work in git like any code.

~/research
● local · 0 egress
CORPUS · ~/research
📁 manuscripts/
📄 hfref-review.md · 3.4 KB
📄 sgtl2-meta.md · 7.1 KB
📁 corpus/
📄 mcmurray2019.pdf · verified
📄 packer2020.pdf · verified
📄 zinman2015.pdf · verified
📁 .intelectra/
📦 qdrant.db · 218 MB
📦 medcpt-article.gguf · 442 MB
📜 prisma.log · 1.2 MB
egress: none
FEATURE · 06

Bibliography Auto-synthesis.

Builds, deduplicates, and gap-analyses a working bibliography against the citation graph it inferred. Vancouver, APA, BibTeX, RIS export.

How it works

The Bibliography Generator agent maintains a structured set of references for the active document. Deduplication happens by DOI then PMID. Gap reports flag under-cited subtopics by year, author cluster, and embedding distance from the centroid of your set.

bibliography.json · live
● 30 papers · 2 gaps
BIBLIOGRAPHY · 30 papers · 4 deduped · 2 retracted excluded
1.
Dapagliflozin in Patients with Heart Failure...
McMurray JJV et al. · NEJM 2019 · PMID 31535829
Verified
2.
Cardiovascular and Renal Outcomes with Empagliflozin...
Packer M et al. · NEJM 2020 · PMID 32865377
Verified
3.
Empagliflozin, Cardiovascular Outcomes...
Zinman B et al. · NEJM 2015 · PMID 26378978
Verified
4.
Liraglutide and Cardiovascular Outcomes...
Marso SP et al. · NEJM 2016 · PMID 27295427
Verified
GAP REPORT · 2 under-cited subtopics
  • Paediatric HFrEF: 0/30 papers
  • SGLT2 in advanced CKD (eGFR < 30): 1/30 papers
FEATURE · 07

Cross-Validation of Findings.

Locate a claim across a corpus; tag each match as direct, partial, indirect, or contradictory. The replication grid surfaces disagreement at a glance.

How it works

The Cross-Validator decomposes your claim into checkable sub-claims, runs each across the corpus, and emits a per-study verdict. The replication grid is exportable as a 1-page evidence card. Quotes are verbatim; verdicts are auditable.

claim · validate
● 9 studies · cross-checked
CLAIM · “metformin reduces all-cause mortality in T2D”
4
Direct support
3
Partial support
1
Indirect contradiction
1
Contextual rebuttal
REPLICATION · per-study
Bannister 14
Griffin 18
UKPDS 34
DPP 02
HOME 09
SPREAD-DIMCAD 13
REMOVAL 17
COSMIC 03
ADOPT 06
FEATURE · 08

Research Profile and Topic Monitoring.

Profile a researcher's prior work, build a topic neighbourhood, and subscribe to PubMed for new matches above a relevance threshold.

How it works

The Research Profiler clusters an author's output by MeSH descriptors and embedding distance. The Literature Monitor uses the resulting profile to set a daily PubMed subscription. Matches above threshold trigger a digest; matches below are logged but suppressed.

Agents
● literature reviewer · active
AGENTS
Literature Reviewer
Bibliography Generator
Cross-Validator
Literature Monitor
Research Profiler
Report Writer
Research Assistant

Compare Intelectra: vs Elicit · vs Consensus · vs Scite