Built so every paper you read compounds.

The AI research agent for citable evidence.

Every paper you read becomes a structured finding in your project. Cited, persisted, ready the next time you ask.

Every finding, citable

From cited finding to the exact line in the source PDF.

Every finding lands in the evidence panel with a typed subject, predicate, object, and the verbatim passage it came from. Open any row and the source PDF jumps to the page, with the matched line highlighted.

EvidenceFull view

Search findings…

51 findings

Group:AllPaper

PD-L1 is expressed in tumor cells in checkpoint inhibitor responders
"...where high PD-L1 expression in tumor cells correlated with longer survival in checkpoint inhibitor responders."
Smith et al. 2024page 4
IL-6 is elevated in CRS patients receiving CAR-T therapy
Wang et al. 2025 · p.7
CD19 CAR-T induced complete remission in B-ALL cohort
Chen et al. 2023 · p.12
Tocilizumab reduced cytokine release syndrome severity
Lopez et al. 2024 · p.3

Smith 2024

Wang 2025

Sourcepage 4 of 12

...where high PD-L1 expression in tumor cells correlated with longer survival in checkpoint inhibitor responders.

From cited finding to the exact line in the source PDF. Every finding carries its passage and page.

Every artifact, one evidence base

Built so the work compounds, not scatters.

Research normally scatters across PDFs in a folder, notes in a doc, comparisons in a spreadsheet, references in a manager. Nothing connects. EvidX keeps every artifact in one project, tied to the same evidence base. The table you built yesterday feeds the draft you write today.

EvidX reading view with the PDF on the left and AI chat with citations on the right

Read PDFs side-by-side with the assistant. Highlights round-trip back to the source.

Learn more about papers

Workspace

One project. Every output.

Papers, sheets, docs, and maps live in the same workspace, with a shared evidence pool. No tab juggling, no re-uploading.

EvidX workspace showing the explorer, an open paper tab, and the agent panel side by side

See how the workspace fits together

Tuned to your field

A vocabulary built around your research.

EvidX learns from your papers, the ones you have published or a few you pick, and designs a typed vocabulary tuned to the entities and relationships you actually study. The home page suggests prompts in that vocabulary, and the copilot extracts every paper into the same fields.

1Onboarding

Confirm your research profile

We read your work as the following. Edit if needed. These drive your home page suggestions.

Areas of interest

biomedical information extractionknowledge graph constructionnatural language processing for biologylarge language models in biomedical researchintracellular signaling pathway modeling

Add another area

+ Add

Looks good, continue

2Academic identity

Linked

University of Pittsburgh

1 linked profile · 8 papers · ORCID linked

Switch

Unlink

Linked publications· 8 total

· Generalizable Biomedical Relation Extraction with LLM Prompting(2024)
· Schema-Guided Event Extraction in Cancer Pathway Literature(2024)
· Cross-Paper Knowledge Graph Construction from Signaling Studies(2023)
· Evaluating Hallucination in Biomedical Relation Extraction Models(2023)
· BioRECIPE: Executable Mechanistic Models from Literature(2022)

3Personalized prompts

Set up projectDiscoverCompareVisualizeDraft

Personalized for youRegenerate

Set up a project on CAR T cell intracellular signaling domains: find 5 foundational papers and extract structured findings

Find 5 recent papers on LLM-assisted biomedical knowledge graph construction, add to a new project, extract findings

Build an evidence base on BioRECIPE and executable mechanistic models with 5 methodologically diverse papers; extract findings

4Copilot

Find 5 recent papers on LLM-assisted biomedical knowledge graph construction, add to a new project, extract findings

Reading 5 papers, extracting along your vocabulary

Hybrid LLM + rule pipelines outperform pure prompting for biomedical relation extraction, with the largest gains on long-tail predicates. BioRECIPE-style typed schemas reduce hallucination on signaling-pathway claims.

Chen 2024, p.4Park 2023, p.7Liu 2024, p.2

22 findings extracted to your information-extraction vocabulary

Sign in, confirm your focus, and the copilot already speaks your vocabulary. Every paper it reads extracts into the same fields.

Read once. Query forever.

EvidX turns every paper into structured evidence that compounds across projects. The next time you ask, the answer is already there, cited.

Built so every paper you read compounds.

The AI research agent for citable evidence.

Every paper you read becomes a structured finding in your project. Cited, persisted, ready the next time you ask.

See how it works

Every finding, citable

From cited finding to the exact line in the source PDF.

EvidenceFull view

Search findings…

51 findings

Group:AllPaper

PD-L1 is expressed in tumor cells in checkpoint inhibitor responders
"...where high PD-L1 expression in tumor cells correlated with longer survival in checkpoint inhibitor responders."
Smith et al. 2024page 4
IL-6 is elevated in CRS patients receiving CAR-T therapy
Wang et al. 2025 · p.7
CD19 CAR-T induced complete remission in B-ALL cohort
Chen et al. 2023 · p.12
Tocilizumab reduced cytokine release syndrome severity
Lopez et al. 2024 · p.3

Smith 2024

Wang 2025

Sourcepage 4 of 12

...where high PD-L1 expression in tumor cells correlated with longer survival in checkpoint inhibitor responders.

From cited finding to the exact line in the source PDF. Every finding carries its passage and page.

Every artifact, one evidence base

Built so the work compounds, not scatters.

Read PDFs side-by-side with the assistant. Highlights round-trip back to the source.

Learn more about papers

Workspace

One project. Every output.

Papers, sheets, docs, and maps live in the same workspace, with a shared evidence pool. No tab juggling, no re-uploading.

See how the workspace fits together

Tuned to your field

A vocabulary built around your research.

1Onboarding

Confirm your research profile

We read your work as the following. Edit if needed. These drive your home page suggestions.

Areas of interest

biomedical information extractionknowledge graph constructionnatural language processing for biologylarge language models in biomedical researchintracellular signaling pathway modeling

Add another area

+ Add

Looks good, continue

2Academic identity

Linked

University of Pittsburgh

1 linked profile · 8 papers · ORCID linked

Switch

Unlink

Linked publications· 8 total

· Generalizable Biomedical Relation Extraction with LLM Prompting(2024)
· Schema-Guided Event Extraction in Cancer Pathway Literature(2024)
· Cross-Paper Knowledge Graph Construction from Signaling Studies(2023)
· Evaluating Hallucination in Biomedical Relation Extraction Models(2023)
· BioRECIPE: Executable Mechanistic Models from Literature(2022)

3Personalized prompts

Set up projectDiscoverCompareVisualizeDraft

Personalized for youRegenerate

Set up a project on CAR T cell intracellular signaling domains: find 5 foundational papers and extract structured findings

Find 5 recent papers on LLM-assisted biomedical knowledge graph construction, add to a new project, extract findings

Build an evidence base on BioRECIPE and executable mechanistic models with 5 methodologically diverse papers; extract findings

4Copilot

Find 5 recent papers on LLM-assisted biomedical knowledge graph construction, add to a new project, extract findings

Reading 5 papers, extracting along your vocabulary

Chen 2024, p.4Park 2023, p.7Liu 2024, p.2

22 findings extracted to your information-extraction vocabulary

Sign in, confirm your focus, and the copilot already speaks your vocabulary. Every paper it reads extracts into the same fields.

Read once. Query forever.

EvidX turns every paper into structured evidence that compounds across projects. The next time you ask, the answer is already there, cited.