Paper Library
PDF upload, DOI/PMID import, metadata enrichment, project collections, tags, and processing states.
DNA Dojo
AI Biology Workspace
DNA Dojo Paper is active first
The AI-native workspace for biology
DNA Dojo turns biology papers into structured research intelligence: citation-backed answers, experimental methods, biological entity memory, public database links, and project-level synthesis.
Citation-backed Q&A
Methods extraction
Project memory
Paper intelligence run
Evidence locked
Paper Card
MAPK rebound after KRAS inhibition
Extracts models, perturbations, controls, readouts, limitations, and source-backed follow-up questions.
MethodsKRAS G12C
Citationsphospho-ERK
Entitiesvehicle control
Citation graph
Answers route through embedded chunks and return citations before they become saved project memory.
Methods p.4Results p.7
DNA Dojo Paper
From PDF to structured research intelligence
The first product surface is built around a complete paper-to-knowledge loop: import, parse, extract, link, ask, synthesize, and save.
Paper Intelligence
Paper Card, Methods Card, Findings Card, Limitations Card, Citation Map, and suggested research questions.
Evidence Q&A
Retrieval-grounded answers with source snippets, page-aware citations, confidence, and insufficient-evidence handling.
Figure/Table Understanding
Extract captions, summarize figures and tables, link visuals back to claims, and store figure assets.
Entity Memory
Genes, proteins, compounds, pathways, diseases, variants, assays, and organisms become searchable memory.
Project Synthesis
Ask across papers, generate evidence matrices, detect contradictions, and create mini-review drafts.
Workflow
Paper to knowledge, with every step traceable
DNA Dojo keeps the interface close to the way researchers work: evidence first, methods visible, entities normalized, and outputs saved back to project memory.
KRAS inhibitor resistance
3 papers · 12 questions · 8 saved cards
Evidence ready
Paper Card
Research question, novelty, core finding, limitations, entities, and follow-up questions.
Methods Card
Model system, perturbation, assay, control group, variables, readouts, and protocol notes.
Citation Q&A
Each answer is generated from retrieved chunks and saved with citation metadata.
Project memory
Paper summaries, methods, answers, figures, and entities become reusable research memory.
1
Ingest
Upload PDFs or import papers by DOI, PMID, or title search.
2
Extract
Parse paper text, figures, tables, methods, claims, and limitations.
3
Link
Normalize genes, proteins, compounds, pathways, diseases, assays, and organisms.
4
Ask
Question one paper or a project library with citation-backed retrieval.
5
Synthesize
Compare findings, surface contradictions, and generate evidence matrices.
6
Remember
Save useful outputs as project cards and reusable biological memory.
Evidence-backed AI
Answers stay attached to source evidence
DNA Dojo is designed for scientific work where every factual statement needs a source trail. Retrieval, citations, snippets, and uncertainty are first-class interface elements.
Citation-backed answer
Which controls did the paper use for the inhibitor assay?
Answer with uncertainty
The paper supports vehicle-treated controls for baseline pathway activity and a combination arm for rebound signaling tests. Evidence outside those retrieved chunks is not assumed.
Methods p.4
Vehicle-treated cells establish baseline pathway activity before inhibitor exposure.
Results p.7
Combination treatment tests rebound signaling after pathway blockade.
Biology layer
Designed around biological evidence, not generic chat
DNA Dojo treats papers as experimental records. The interface privileges methods, controls, model systems, entities, citations, uncertainty, and reusable project memory.
Entity memory preview
KRAS
ERK1/2
Sotorasib
phospho-ERK
lung adenocarcinoma
RTK signaling
Paper-grounded synthesis
Research questions, claims, methods, entities, caveats, and suggested follow-ups stay tied to source chunks.
Experimental method extraction
Model systems, perturbations, assays, controls, readouts, and protocol notes are normalized into reusable cards.
Biology memory layer
Saved cards turn papers and answers into project memory for genes, pathways, compounds, and assays.
Public database layer
Connect paper intelligence to trusted biology resources
Phase 1 extends beyond PDF parsing by linking extracted entities and metadata to open literature, protein, genome, compound, and structure databases.
Database linking layer
PubMed
Literature
UniProt
Protein
Ensembl
Genome
ChEMBL
Compound
Normalized entity
KRASQ61HMAPK signaling
PubMed / NCBI
Literature and genesPMID metadata, abstracts, PubMed records, PMC links, and Entrez database references.
Europe PMC
Open full textLife-science metadata, open-access full text, figures, and supplementary material when available.
OpenAlex
Scholarly graphWorks, authors, sources, institutions, topics, and citation-network context.
UniProt
Protein annotationProtein names, functions, organisms, accessions, and cross-references.
Ensembl
Genome annotationGenes, transcripts, variants, sequences, homology, and regulatory context.
ChEMBL
Compound evidenceMolecules, targets, assays, mechanisms, and bioactivity measurements.
RCSB PDB
Experimental structureProtein and complex structures, ligands, chains, assemblies, and structure metadata.
AlphaFold DB
Predicted structurePredicted protein structures and confidence context for sequence-to-structure workflows.
Use cases
Built for serious biology research loops
The product should help users move from reading to evidence extraction, from evidence to mechanisms, and from mechanisms to experiment planning.
Read papers faster
Turn dense biology papers into structured questions, findings, methods, and limitations.
Extract methods
Capture models, perturbations, assays, controls, time points, readouts, and protocol notes.
Map mechanisms
Connect genes, proteins, compounds, pathways, phenotypes, and diseases into project memory.
Plan experiments
Convert paper evidence into constraints, controls, next assays, and follow-up hypotheses.
Build literature matrices
Compare multiple papers by claim, evidence, organism, method, and citation strength.
Prepare reviews
Generate project-level synthesis with citations and clearly marked uncertainty.
Platform
A workspace that can grow beyond papers
Paper is live first. Molecule, sequence, experiment, tutoring, and discovery modules are reserved so future tools can share the same project memory layer.
Paper
Active in V1Read, query, and card biology papers with citations.
Molecule
ReservedTrack compounds, targets, mechanisms, and assay evidence.
DNA / Protein
ReservedAnnotate sequences and connect variants to literature.
Tutor
ReservedTurn project context into guided explanations and checks.
Experiment
ReservedDraft experiment plans from methods and constraints.
Discovery
ReservedSurface hypotheses, gaps, and next papers to read.
Security
Private by design, evidence-first by behavior
The research workspace needs to be careful with user papers, project memory, AI keys, and scientific claims.
Private projects and authenticated workspaces
Supabase row-level security for user-owned data
Server-side OpenAI and Supabase service keys
Citation-first answers with source snippets
Usage events and AI spend guardrails
Recoverable processing states and audit-friendly analysis runs
Final CTA
Start with DNA Dojo Paper
Build a private biology workspace where uploaded papers become citation-backed cards, methods intelligence, entity memory, and project synthesis.