ARCHITECTURE

Local-first, method-transparent

Nucleora runs entirely on your machine — your sequences never leave your computer. Here's how the pieces fit together.

ARCHITECTURE

A local web app

Nucleora is a Python application that runs a local web server on your machine. When you launch it, it opens in your browser — but everything happens locally.

  • Flask backend handles API requests
  • Browser frontend provides the interface
  • localhost:5000 — nothing leaves your network
  • No account required — no cloud, no login
[Your Browser]
↓ ↑
localhost:5000
↓ ↑
[Python Backend]
Everything runs on your machine
COMPUTATION

Real algorithms, not approximations

Nucleora uses established bioinformatics libraries — the same ones you'd use in a Python script or a published paper.

  • ViennaRNA for RNA secondary structure prediction
  • BioPython for sequence parsing and analysis
  • primer3 for primer design
  • REBASE restriction enzyme database

Every calculation traces to a named, published algorithm — not a black box.

ViennaRNA
MFE folding, ensemble, BPP
BioPython
Sequence I/O, translation, alignment
primer3
Tm, dimers, hairpins
DATA

Real codon tables, documented sources

Codon optimization is only as good as the codon-usage data. Nucleora ships tables built from NCBI RefSeq coding sequences — not inherited from generic "mammalian" approximations.

  • RefSeq-derived tables for well-sequenced species
  • Clade proxies for species with limited data — explicitly flagged
  • Gene count shown for each table
  • No silent fallbacks — you always know the data quality
Bos taurus
RefSeq · 42,847 CDSs
Loxodonta africana
RefSeq · 103,241 CDSs
Okapia johnstoni
Proxy: Bos taurus · 1 CDS
NETWORK

Optional database access

The only network requests Nucleora makes are database lookups that you initiate — searching for an antigen, fetching a sequence by accession.

  • NCBI Entrez for nucleotide/protein searches
  • UniProt for protein annotations
  • Ensembl for gene/transcript data
  • PDB/RCSB for structure references

Results are cached locally. Core computation — folding, optimization, primer design — works fully offline.

[Your query: "spike glycoprotein"]
NCBI / UniProt / Ensembl
[Cached locally]
Your sequences are never uploaded
GETTING STARTED

Simple setup, no command line

When Nucleora becomes available, it will use a double-click launcher that handles everything automatically — no terminal commands, no manual configuration.

  • Self-contained — manages its own Python environment
  • Cross-platform — same experience on macOS and Windows
  • Auto-updating — stays current without manual steps

Request early access to be notified when builds are available.

Coming soon
macOS & Windows
Double-click launcher
No command line required

Ready to try it?

Request early access and we'll reach out when builds are available for your platform.