Platform

The enzyme design engine.

From reaction specification to ranked, expression-ready sequences in one request.

How the pipeline works

Reaction Input EC Classifier Structure Retrieval Generative Decoder Biophysical Scoring Expr · Solub · Tm Ranked FASTA

EC Classification

Your substrate/product SMILES or enzyme class description is mapped to EC hierarchy via a fine-tuned reaction classifier, providing structural context for generation.

Generative Decoder

An autoregressive transformer decodes novel sequences conditioned on reaction embeddings. Output is not retrieval — these are sequences outside any training homolog cluster.

Biophysical Scoring

Parallel prediction heads score each candidate for E. coli expression yield category, solubility probability, and thermostability Tm range before returning ranked results.

One endpoint. Full pipeline.

POST your reaction context and receive ranked, expression-ready sequences in one call. The Python SDK wraps the REST API with async support and result polling built in.

Full API reference →
Python
import fermvyne

client = fermvyne.Client(api_key="fvyn_...")

result = client.design(
  substrate="CC(=O)Nc1ccc(O)cc1",
  product="CC(O)Nc1ccc(O)cc1",
  targets={
    "tm_min": 60,
    "host": "ecoli",
    "n_candidates": 12
  }
)

for seq in result.candidates:
  print(seq.fasta, seq.tm_estimate, seq.expression_score)
Response (JSON)
{
  "job_id": "j_8xkp42",
  "candidates": [
    {
      "rank": 1,
      "fasta": "MKVLSGEDK...",
      "tm_estimate": 67.4,
      "expression_score": "high",
      "solubility_prob": 0.91
    }
  ]
}

Connects to your existing lab stack

Benchling ELN

Push enzyme design results directly into your Benchling notebook. FASTA sequences and prediction scores appear as structured data entries.

Python SDK

Install with pip. Full async support, automatic result polling, and pandas-compatible DataFrames for downstream analysis in your existing notebooks.

REST API

Language-agnostic JSON endpoints. Design requests, result retrieval, and pathway co-design all accessible with a single API key from your dashboard.

FASTA Export

Direct FASTA downloads formatted for Twist Bioscience and IDT gene synthesis ordering. One click from ranked candidates to synthesis order.

Start free. Scale when it works.

Free

$0 /month

No credit card. Run Fermvyne on your own enzyme targets — validate Tm predictions against your DSF data before committing to a paid plan.

  • 10 designs per month
  • Basic expression + solubility predictions
  • FASTA download
  • Web interface
Get started free

Lab

$799 /month

For labs running multiple concurrent campaigns. Pathway co-design endpoint unlocked. Benchling ELN sync so results land directly in your experiment entries.

  • Multi-user (up to 8 seats)
  • Pathway co-design (4+ enzyme routes)
  • Benchling ELN sync
  • Priority support + onboarding
Contact for Lab

Run your first design request.

No wet lab required. Results in under 10 minutes.