Scientific computing for the life sciences.

Design, simulate, analyze and verify across DNA, proteins, structures and chemistry, in one environment, on one engine validated against the reference implementations of each field, open to scientists and to their AI agents alike.

Open the workbench Browse every tool

Runs offline
Golden-validated
No account
Free, whole

Pick a tool

Seven ways in. One editor.

Cloning simulationGibson, Golden Gate, Gateway, MoClo

Open the workbench

The whole argument, on one result

Every number opens.

One real result from the qPCR panel, shown at every depth it offers. Nothing here is a summary of the picture beside it: the picture is the product, running.

FIRST
A sentence, not a number
The panel writes its own conclusion from the run it just did: which thing moved, by how much, and with what interval around it. It is generated from the result, so it cannot describe a run that did not happen.
THEN
What would make it wrong
Every assumption the method rests on is stated next to the answer, and each one names its own subject rather than warning in general. An assumption nobody tested says so, and says which control would test it.
UNDERNEATH
The machinery, in one line
The model, the estimator, the reference it normalized to, the calibrator it compared against, and the published method it follows. Nothing is a house convention: the panel names the paper.

Open the workbench

What you can get done

You came here with a construct to build, not a tool to open.

36 things people actually come here to do. There is no panel called “design a construct”: there are several that together do it, and each card links the ones that do, so you can check the claim before you trust it.

Design

Draw the construct, and every oligo and edit it needs.

Design a plasmid out of the parts you already haveGibson Assembly Golden Gate Assembly Grammar Assembly Ligation
Plan a multi-part build that depends on itselfAssembly Plan Grammar Assembly
Get a gene to express in a new hostCodon Optimization Codon Usage Analysis Codon Tables
Tune how strongly that gene is translatedRBS & Operon Design
Order primers that work the first timePrimer Designer Melting Temperature Primer Library
Change one residue, or remove one siteSite-Directed Mutagenesis
Build a variant library for protein engineeringSite-saturation mutagenesis Multi-site mutagenesis
Knock out or edit a gene with CRISPRCRISPR Guide Design
Clone a gene nobody has sequenced yetCODEHOP hybrid primers from a protein alignment Degenerate primers from an alignment
Build a field diagnostic that needs no thermocyclerLAMP primer set
Measure methylation at a locusBisulfite methylation primers
Image one transcript inside a cellIn-Situ Probes
Pool many samples onto one sequencing runBarcode & Index Design
Make a cell compute a logic functionGenetic Circuit
Express and purify a protein, tag and allBack-translate: protein → DNA, codon-usage weighted or degenerate Protein Analysis Protease Digest

Simulate

Run the bench step before you run it at the bench.

Know what your digest will look like before you run itRestriction Digest Virtual Gel Bench Enzyme Collections
Check a primer pair against the template it will meetPCR Simulation Fusion PCR
Stop a multiplex failing on the pair nobody checkedMultiplex PCR
Run a Gateway or recombinase reaction on paper firstRecombination Cloning
Run the same chain on a hundred constructsWorkflow Builder

Analyze

Read what came back, at whatever depth the question needs.

Turn a qPCR plate into numbers you can defendqPCR Design Amplification Curves → Cq Melt Curves → Tm
Get an absolute concentration out of a ddPCR runDroplet gating → concentration Droplet classification Digital PCR
Read a Sanger trace properly, not just its basesChromatogram Fragment analysis
Work out what your CRISPR edit actually producedCRISPR Edit Inference
Place a family of sequences on a tree you can trustMultiple Sequence Alignment Phylogenetics EMBL-EBI Services
See the structure an alignment score hidesDot Plot
Find out what a sequence is, and where else it occursBLAST Search Search Databases Protein Domains
Annotate a plasmid that arrived as bare basesAuto-Annotate Apply features from a file
Read a cohort’s variants population by populationPopulation genetics: allele frequencies, Hardy–Weinberg, Fst, linkage disequilibrium Variant Calls
Identify proteins from a mass-spec runMass spectrometry: peptide masses, b ladders, isotopes, spectrum scoring, q-values, protein inference
Characterize an antibody the way the literature doesAntibody Numbering
Keep the paper, the structure and the spreadsheet with the constructStructure Molecule Image Table

Verify

Prove the thing you built is the thing you drew.

Prove the clone you got is the construct you drewSequencing Verification Map reads to a reference
Find the vector contamination before a reviewer doesScreen for vector contamination
Recover how a construct was made, years laterHistory
Write the methods section while the work is freshLab Notebook History

What's inside

Everything on one canvas, reading one project

All 57 live on one canvas and read the same project, so a result in one is an input to the next: design, simulate, analyze, verify.

Annotate & designRestriction digestPrimer designCloning simulationRecombination cloningPCR simulationqPCR design & analysisSanger tracesAlignment viewer (BAM)Multiple sequence alignmentPhylogeneticsTranslate & ORFsCodon optimizationUniversal file viewerCRISPR guide designSite-directed mutagenesisGrammar assemblyRBS & operon designBarcode & index designMelting temperatureCodon usage & CAICodon-table builderEnzyme collectionsVirtual gel benchPrimer librarySequencing verificationHistory & provenanceLab notebook (ELN)Genetic circuitsBench calculatorsPower and sample sizeTrade-offsIn-situ probesRNA-seqMass spectrometryProtein domainsAntibody numberingBLAST searchMap readsChromosomal microarrayFlow cytometryFragment analysisDot plotFetch from databaseLigationMulti-level assembly plansAuto-annotateBack-translationCRISPR edit inferenceContamination screenProtein analysisEMBL-EBI servicesVariant callsPopulation geneticsSmall-molecule assessmentDiscovery analysesWorkflow pipelines

Bring your files

The sequence is the easy half.

Open the file you already have and the build comes with it: every step, every input, every parameter, readable on screen and back out again in a format that is not ours.

30 of 31history-bearing GenBanks in QUEEN’s own demo tree, decoded and checked against real QUEENThe longest is a 76-step, 12-part Golden Gate build.

This is a SnapGene .dna file opened as it came, with nothing typed in. The badge names the record the decoder found; the counts are read off that record; the sentence under them is written from this decode and not from a template, so a different file says something different. Everything the two columns below promise is on this screen or one tab away from it.

Read from

SnapGene .dnathe reaction graph the file embeds, walked as a graph rather than skimmed as a list
QUEEN GenBankthe build script the file carries in its own annotations, decoded back into operations
OpenCloning strategythe interchange document, read back into the reactions it describes
SBOL documentthe PROV-O activities the document records, read back into the steps they describe
Built herethe log this workbench writes as its own cloning tools run

Written back out

GenBank with construction historythe build travels inside the molecule
OpenCloning document (JSON)the whole strategy, loadable in pydna
Reconstruction script (Python)the build as code that runs and rebuilds it
Methods (Markdown)the numbered steps, ready for a manuscript
Steps (CSV)one row per operation, for a notebook or a sheet
Full record (JSON)everything the decoder produced, re-checkable

Operating manual

Everything the editor does, on one page.

4 clauses

§1

Scope

Jean is a scientific computing environment for the life sciences. It covers the work a bench scientist does at a computer in a given week, annotating a plasmid, designing a primer, planning a digest, simulating a clone, fitting a curve, reading a trace, in one program rather than eleven, and it shows its working at every step.

57 surfaces are listed at §3.1 and 24 engines at §3.2. The libraries those engines derive from are registered at §4.

§2

Operation

2.1Environment

Runs in a browser tab. No installation, no runtime to provision, no license server to reach, and no account to create before a file can be opened.

2.2Residency

Files are read from your disk and stay there. Nothing is uploaded, because there is no endpoint to upload to.

2.3Cost

Free, with no tier held back.

§3

Instruments

01Design

Draw the construct, and every oligo and edit it needs.

02Simulate

Run the bench step before you run it at the bench.

03Analyze

Read what came back, at whatever depth the question needs.

04Verify

Prove the thing you built is the thing you drew.

One record, moved through the four stages. Every value above was computed in this tab from data/samples/Genbank/PBR322.gb · GenBank SYNPBR322, by the same parser and the same cut-site finder the editor runs.

3.1Editor surfaces (57)

Each row is a live surface in the editor, not a planned one, and each sits under the stage of the arc it is reached for from.

Design20

01Annotate & designLinear + circular maps, features, parts

02Primer designTm, dimers, hairpins, product size

03Codon optimizationA real constraint solver, not a lookup table

04CRISPR guide designgRNA scoring, off-target scan, edit outcomes

05Site-directed mutagenesisQuikChange / SOE primers + silent site scrub

06Grammar assemblyMoClo & GoldenBraid standard parts

07RBS & operon designSalis-style SD·anti-SD strength + operon builder

08Barcode & index designDistance-checked sets, color balance

09Melting temperatureShared Tm conditions every tool designs against

10Enzyme collectionsNamed sets that scope your digest

11Primer libraryOligo lifecycle + order / IDT export

12Genetic circuitsLogic from sensors, gates and reporters

13Bench calculatorsDilutions, serial dilutions, mass and moles, master mixes

14Power and sample sizeReplicates an experiment needs, including RNA-seq

15Trade-offsCandidates no other beats on every objective

16In-situ probesTiled, padlock, HCR, MERFISH and Xenium probes

17Fetch from databaseNCBI & ENA search → open the record

18Multi-level assembly plansPlan a build across assembly levels

19Auto-annotateFind known features on a new sequence

20Back-translationProtein to DNA, codon-weighted or degenerate

Simulate7

21Restriction digestCut maps, fragments & a simulated gel

22Cloning simulationGibson, Golden Gate, Gateway, MoClo

23Recombination cloningGateway, Cre-Lox, integrase, CRISPR knock-in

24PCR simulationAmplify, fusion, multiplex & nested

25Virtual gel benchMulti-lane agarose gel simulator

26LigationSticky, blunt, and cut-and-ligate in one step

27Workflow pipelinesWire tools output→input as a DAG, run headless

Analyze24

28qPCR design & analysisProbes, standard curves, ΔΔCt, reference genes

29Sanger tracesOpen .ab1, trim, call peaks, verify

30Alignment viewer (BAM)Indexed pileup; open a read as sequence or MSA

31Multiple sequence alignmentProgressive or partial-order, with a live consensus

32PhylogeneticsDistance trees, bootstrap support, Newick

33Translate & ORFsSix-frame translation, ORF scan, GC

34Universal file viewerAuto-detect → view, convert, analyze

35Codon usage & CAIAdaptation index, rare codons, weights

36Codon-table builderGenetic-code chart from real CDS

37RNA-seqDifferential expression, over-representation, GSEA

38Mass spectrometryPeptide masses, spectra, protein coverage

39Protein domainsInterProScan families + linear motifs

40Antibody numberingChains, CDRs, germlines, species

41BLAST searchYour own project records, searched locally

42Chromosomal microarrayCopy number, LOH and mosaicism on a karyotype canvas

43Flow cytometryOpen .fcs, compensate, gate and count populations

44Fragment analysisOpen .fsa, size peaks against a ladder

45Dot plotSelf and pairwise sequence comparison

46Protein analysisMW, pI, GRAVY, extinction and protease digests

47EMBL-EBI servicesClustal Omega, MUSCLE, needle, water, phmmer, Phobius

48Variant callsVCF and BCF: counts, Ts/Tv, per-sample genotypes

49Population geneticsAllele frequencies, Hardy–Weinberg, Fst, LD, PCA

50Small-molecule assessmentDrug-likeness rules, structural alerts, descriptors

51Discovery analysesCodon outliers, tree incongruence, assay decay, Discovery Expert

Verify6

52Sequencing verificationGrade Sanger / NGS reads vs your construct

53History & provenanceLineage tree + bench protocol

54Lab notebook (ELN)Experiments with clickable @plasmid chips

55Map readsReads and contigs onto your reference

56CRISPR edit inferenceIndel spectra from edited and control Sanger reads

57Contamination screenVecScreen against UniVec, plus masking

3.2Headless engines (24)

Framework-free logic, callable without the interface, validated per §4.

Restriction & digestionREBASE enzymes · cut sites · fragments

Assembly planningGibson · Golden Gate · BioBrick · BASIC · LCR

PCR · fusion · multiplexamplify · overlap-extension · primer screen

Primer design & screendesign · assembly fragments · overlaps

Melting temperatureSantaLucia NN + Owczarzy Mg²⁺

Secondary structureprimer3 thal ΔG · hairpin · homo/hetero-dimer

Nucleic-acid foldingZuker MFE: seqfold DNA + ViennaRNA-model RNA

Codon optimizationDnaChisel constraint + objective solver

RecombinationGateway · Cre-Lox · integrase · HR · CRISPR knock-in

CRISPR guide scoringon-target scores + cas-offinder off-target search

Site-directed mutagenesisQuikChange / SOE design + silent-site scrub

RBS & operon designSD·anti-SD hybridisation strength, folded

Gel electrophoresismigration model + per-lane band profiles

Multiple sequence alignmentpartial-order graph align + MSA scoring

PhylogeneticsBIONJ distance trees · bootstrap · consensus

Barcode set designdistance-checked sets · color balance

qPCR design & analysisprobes · standard curves · ΔΔCt · reference genes

Fuzzy searchsequence + feature, mismatch-tolerant

Provenance & historyQUEEN quine · build graph · diff

Assembly reportsconstruct sheets · simulation rows · mutation counts

Plasmid map renderingpublication-ready circular/linear SVG

Sanger analysisconsensus · mutation calls · allele phasing

GenBank · FASTA · SnapGeneparse, repair & write, round-trip clean

BAM alignmentsBGZF · BAI/CSI indexed random access

§4

Provenance

No engine here is a port of a library. For each capability the published implementations are studied, the best of each is merged into one engine, and our output is then asserted byte for byte against every one of them, including the cases where they disagree with each other. Each source is cloned to disk and executed for real, so the vectors can be re-derived rather than trusted.

Library	Language	Covers
pydna	Python	assembly · amplify · gel · Dseq
DnaCauldron	Python	Golden Gate · combinatorial assembly
QUEEN	Python	provenance · quine · build graph
DnaChisel	Python	codon optimization · constraints
SeqTrace	Python	Sanger trim · consensus
cfutils	Python	trace ↔ reference mutation calls
sangerseqR	R	mixed peaks · allele phasing
poapy	Python	partial-order MSA
pyMSA	Python	alignment scoring
crisprScore	R	guide efficiency scoring
cas-offinder	C++	off-target search
sgffp	Python	GFF parsing
@gmod/bam	JS	BGZF · BAI/CSI random access
primer3	C	thal ΔG · hairpin · dimer
seqfold	Rust	Zuker DNA folding
SpliceCraft	Python	RBS · mutagenesis · map image
ape	R	BIONJ phylogenetic trees
DNABarcodes	R	barcode set design

Coverage

Validated against, by language.

Python10 · 56%

R4 · 22%

C++1 · 6%

JS1 · 6%

C1 · 6%

Rust1 · 6%

Every one of the 18 libraries at right is cloned, run for real, and asserted against: the source language is whatever the original authors used.

21of 86 formats

▾ GenBank

The extension routes it and the contents confirm it: drop any of the 86 formats in and it just opens.

The engine, in the round.

headless · golden-validated

Universal viewer86 formats

Cloning simulationGibson · Golden Gate · MoCloTwo fragments, 28 bp homology. Assembled and checked against pydna.

Primer designTm, dimers, hairpins

Sanger tracesopen, trim, call peaks

Mass spectrometrypeptides to proteins

Checked to the byte.the reference implementations run for real, our output asserted against theirs

Alignment viewerindexed BAM, random access

Codon optimizationa real constraint solver

ATGCTGAAAGGCTAA

CRISPR guidesscoring + off-target scan

Uploaded this session0 bytesNo network calls

Validated against18/18

every one asserted byte for byte

Surfaces and engines81

57 editor24 engine

Every surface, one cell each

blue: editor surfaces · green: headless engines

Opens on sight

GenBankannotated sequence

FASTAsequence, unannotated

FASTQreads with per-base quality

EMBLannotated sequence, ENA

GFF3feature annotations

GTFgene transfer annotations

+15 more, no import step

Marginalia

On showing the work

A number on its own is a claim. Every result here carries the model that produced it, the assumptions it stands on, and the reason it was flagged or trusted.

Read on

On being checked

We did not port a library. We studied every serious implementation of each capability, merged the best of them into one engine, and assert our output against all of them.

Read on

On twenty-seven genes

Excel kept turning gene names into dates, so the field renamed the genes. A spreadsheet that knows what SEPT2 is costs nothing to build.

Read on

Pick one up

Restriction digest Cloning Primers Sanger Spreadsheet Notebook Alignment CRISPR qPCR Gels Phylogenetics Provenance

Open wide.
Work close.
Keep it.

An environment where the tools know about each other, every default is visible, and nothing is held back for a paid tier.

Open the workbench

86 formats

It opens anything

Drag in a GenBank record, a trace, an indexed BAM. It works out what it is and opens it.

24 engines

Checked byte for byte

Every engine is checked against the published implementations, byte for byte.

no account

It stays on your disk

No sign-up, no sync, no telemetry. The file stays on the disk you opened it from.

Checked against the libraries the field already trusts.

three of 18

checked againstpydna(assembly, amplify, gel)

Python Golden-validatedbyte for byte

engine:primer3(thal, hairpin, dimer)

ΔG in the browserno server

engine:SeqTrace(trim, consensus)

and sangerseqR for mixed peaksallele phasing

For your assistant

The model does the reasoning. The engine does the science.

A model with a terminal can run anything and still answers a scientific question from what it remembers. Execution is not an oracle. Point an MCP client at /mcp and the model calls the same tools these panels run, each checked against the reference implementation of its field before it ships, and each answer says how it was reached.

Beside this are three calls to the restriction-site tool an assistant gets, run just now in this page by the same code the server runs. The only EcoRI site in this circle spans the point where it closes, so reading it as a line finds nothing: the plausible answer, and the wrong one. A name that resolves to no enzyme is refused by name, because an empty list would read as no sites.

Connect an assistant

It works

the way

you do

01

It opens what you have

Not a subset. The formats a working week actually produces, including the proprietary one.

02

It shows the working

Every result names the method and the parameters it used, and lets you change them.

03

It stays where you put it

A project is a folder on your disk. Move it, back it up, open it on another machine.

04

It does not hold an instrument back

Every tool is in the free plan. The paid one adds a cloud behind it, never a feature you were missing.

01

Viewer

Drop in almost any bio file. It detects the format and just opens.

02

Editor

Not a preview pane. Annotate, simulate, verify, and undo any of it.

gibson_assembly.dna

gibson_assembly.dnadata/samples/SnapGene/gibson_assembly.dna

03

Engine

Headless, framework-free, and checked against the libraries it came from.

Build log

What shipped. What's next.

The same three states every capability in this app moves through, taken straight from the project's own working notes.

Shipped6

Restriction digestCut maps, fragments & a simulated gel

Barcode & index designDistance-checked sets, color balance

Recombination cloningGateway, Cre-Lox, integrase, CRISPR knock-in

Alignment viewer (BAM)Indexed pileup; open a read as sequence or MSA

qPCR design & analysisProbes, standard curves, ΔΔCt, reference genes

Site-directed mutagenesisQuikChange / SOE primers + silent site scrub

Building now4

UI migrationEvery dialog and panel, moving off the old Blueprint kit onto shadcn.

Canvas performanceFaster rendering for very large plasmid and genome maps.

Tool taxonomyRefining how tools are grouped and how dense each panel reads.

Sharing & permissionsShared projects with per-user roles, for the cloud plan.

On the roadmap8

The engine, publishedpackages/ports as an installable library other tools can build on.

RO-Crate exportA complete, shareable research package (no server required).

SBOL 3A free, browser-native tool for the synthetic-biology exchange format.

Live collaborationReal-time, multi-cursor editing on the same step log.

True offline installA desktop build for a machine that never sees a network.

Teaching modeA guided layer over the real editor, for the classroom.

Expression designDescribe the protein and the host; get ranked sequence variants with the reasoning for each.

One request, a whole buildGuides, template, screening primers and assembly steps as one package you can order.

Read the full roadmap

Pick your plan

The whole workbench$0 / forever

All 88 tools, and the 24 engines under them
Checked against the reference implementations, byte for byte
No account, no telemetry, no upload

no tier held back

Editor

57 surfaces on one canvas

Engine

Checked against the reference implementations

Formats

86 read · 29 written

Your disk

Files stay where you put them

Shared projects

Cloud plan, by permission

Agents

Cloud plan, in the cloud

Open the workbench

If something here is wrong, or missing, or you want a format we do not read yet, say so at wordrobesapp@gmail.com

Talk soon.

Open the workbench Read the story

86 formats, detected automaticallyDrop a file in. The app reads its content, works out what it is, and opens it: no import dialog and no conversion step.

AllSequenceTraceAlignmentTabular

SUPPORTED

GenBankannotated sequence

SUPPORTED

FASTAsequence, unannotated

SUPPORTED

FASTQreads with per-base quality

SUPPORTED

EMBLannotated sequence, ENA

SUPPORTED

GFF3feature annotations

SUPPORTED

GTFgene transfer annotations

SUPPORTED

BEDgenomic intervals

SUPPORTED

VCFvariant calls

SUPPORTED

SnapGeneannotated sequence, proprietary

SUPPORTED

Serial Clonerannotated sequence, keeps its colors

SUPPORTED

Swiss-Prot / UniProtannotated protein, UniProtKB

SUPPORTED

SBOLsynthetic-biology parts, XML

SUPPORTED

JBEI / raw sequencesequence, JBEI-SEQ XML

SUPPORTED

Geneioussequence, Geneious XML

SUPPORTED

GenPeptannotated protein

SUPPORTED

AB1 traceSanger trace, Applied Biosystems

SUPPORTED

SCF traceSanger trace, standard chromatogram

SUPPORTED

ZTR traceSanger trace, compressed

SUPPORTED

BAMaligned reads, indexed

SUPPORTED

Parquetcolumnar tabular data

SUPPORTED

CSV · TSVtabular data, gene-safe

Every tool, one page each

All 88 tools

Scientific computing for the life sciences.

Seven ways in. One editor.

Every number opens.

A sentence, not a number

What would make it wrong

The machinery, in one line

You came here with a construct to build, not a tool to open.

Design

Simulate

Analyze

Verify

Everything on one canvas, reading one project

The sequence is the easy half.

Everything the editor does, on one page.

Scope

Operation

Instruments

Provenance

Validated against, by language.

The engine, in the round.

On showing the work

On being checked

On twenty-seven genes

Open wide.Work close.Keep it.

It opens anything

Checked byte for byte

It stays on your disk

Checked against the libraries the field already trusts.

The model does the reasoning. The engine does the science.

It opens what you have

It shows the working

It stays where you put it

It does not hold an instrument back

Viewer

Editor

Engine

What shipped. What's next.

Pick your plan

Editor

Engine

Formats

Your disk

Shared projects

Agents

Open wide.
Work close.
Keep it.