HOW IT WORKS
Five layers. One pipeline.
Nucleora is a five-layer CAD stack built entirely for nucleic acid design. Each layer has a clear job. Together they produce a verifiable, releasable artifact - not just a sequence.
Canvas
Schematic capture - parametric assembly
You assemble a construct from parametric parts - 5′ UTR, Kozak, CDS, 3′ UTR, poly(A) - not by typing characters. Change poly(A) from 80 to 120 nt and the full sequence regenerates underneath you.
The sequence is a rendering of the design. It is never its source of truth. This is how every other mature engineering discipline works, and it is how Nucleora works.
CONSTRUCT ASSEMBLY
RENDERED SEQUENCE (3,897 nt)
Sequence is a rendering. Editing the parts rerenders it.
Library
Versioned, provenanced parts
Every part carries its provenance - the database it came from, the accession number, the date it was fetched. Codon usage tables are tallied from real RefSeq CDS sequences.
When a species doesn't have its own table - an okapi, a bongo, a white rhino - Nucleora names the proxy it used and the distance metric. It never approximates silently.
Constraint Engine
Design rules - continuous, real-time
Design rules run continuously as you design. They check GC content, homopolymer runs, secondary structure at the 5′ UTR, restriction site conflicts, synthesis feasibility, and more.
When a rule fails, Nucleora explains why in plain language and - for many rules - offers a synonymous substitution that resolves the problem without changing the protein.
Simulator
Physics in the loop - local, private
ViennaRNA runs locally. Watch the secondary structure change as you edit. The 5′ UTR fold ΔG updates in real-time with every nucleotide change.
Walk the template through the IVT run before it costs anything: predicted yield, cap incorporation efficiency, manufacturability grade, and an itemized order form - all computed on your machine, nothing sent anywhere.
Release Path
From intent to auditable artifact
When you're satisfied with the design, you cut a release. Nucleora generates one signed, immutable artifact: the design revision, every check result, every waiver, every simulation output, and the bill of materials.
SBOL export. 21 CFR 11 signatures. The full audit trail - in a format another person can verify without running Nucleora at all.
QUESTIONS - FAQ
Common questions.
No. ViennaRNA fold, codon tables, DRC, and all 41 design tools run entirely on your machine. An internet connection is needed only for RefSeq lookups and the initial download. After that, Nucleora works air-gapped.
Yes. Local-first means no sequence leaves your infrastructure. Nucleora makes no outbound connections during normal operation. The 21 CFR 11 audit trail and signed release records are designed for regulated workflows.
Nucleora names the closest proxy it found, the distance metric used, and the n-count of the proxy table. It never approximates silently. You always know exactly what data is backing your design.
Nucleora is in active development. Join the early access list at nucleora.org and we'll reach out when your spot is ready. No account required to request access.
Drag-and-drop part assembly, bidirectional sequence sync, mRNA linear view, IVT plasmid circular view, parametric poly(A) tails, golden-gate and Gibson cloning site design, and automatic re-annotation on sequence change.
41 tools across categories: Cloning (Golden Gate, Gibson, restriction), Codon tools (optimization, tester, lab), Mutagenesis (site-directed, HDR donor designer, prime/base editing, multi-site), Parametric Design, and Parts registry with SBOL2/3 export and SynBioHub sync.
TRY IT
Paste a sequence. See the checks fire.
Runs entirely in your browser — no data leaves this page.
EARLY ACCESS