Solutions
For academic researchers
Pathway engineering tools that a lab can actually cite. Every run keeps its inputs, parameters and engine versions, and everything leaves in formats the rest of your software already reads.
Free
CPU engines
Citable
run identifiers
65
host models
Open
tool stack
01
The problem
The computational half of a plant metabolic engineering project usually lives in a personal environment: a COBRApy notebook that only runs on one laptop, a folder of FASTA files with names that made sense in March, model weights that were current when the analysis ran and have since moved on. The science is fine. The record is not, and the record is what a reviewer, a successor or a future you actually needs.
The second problem is access. The tools that matter now — protein language models, structure prediction, docking — want GPUs, and a group without a cluster allocation either waits or does without.
02
What the platform does about it
Runs, not results
Every computation is stored with its inputs, its parameters and the pinned engine version that produced it. Re-run it a year later and you are re-running what actually ran, not its successor.
Stable identifiers
Runs and construct versions carry identifiers a methods section can point at. 'We ran FBA' becomes a reference to a specific solution with its assumptions attached.
GPU without a cluster
ESM-2 scanning, ESMFold, Boltz-2 and docking run on prepaid credits — $25 for 250 credits, one credit for roughly one Standard job. No allocation request, no queue politics.
Tools you already trust
COBRApy and Escher, Meta ESM-2 650M and ESMFold, AlphaFold and Boltz-2, CHOPCHOP- and CRISPOR-style guide scoring with Cas-OFFinder, DNA Chisel, Primer3, Biopython, pydna, MAFFT, Mol*.
Standard formats out
GenBank, FASTA, SBOL and ST.26 export, plus CSV for anything tabular. Nothing is trapped, which matters when a collaborator has their own pipeline.
Shareable without explanation
A run link carries the inputs, the versions and the output together, so a co-author can review a result without a call to explain what the figure meant.
03
A representative project
- Group
- Plant secondary metabolism lab
- Question
- Pathway transfer across Solanaceae hosts
- Hosts compared
- 48, via one batch
- Flux engine
- COBRApy — CPU, free
- Variant work
- ESM-2 650M on PAMT and pun1 product
- Alignments
- MAFFT across host homologs
- Output
- CSV table, maps, citable run IDs
A comparative question answered as a batch rather than as forty-eight separate sessions. The feasibility sweep cost nothing because flux analysis is CPU work; the credits went where they were actually needed, on scanning the two enzymes that the flux maps had already identified as controlling.
The output a group like this cares about is not the table. It is that the table came from a batch with a manifest, that each row links to a run, and that the whole thing can be handed to a reviewer who asks how it was produced.
04
Getting started
Create an account and the CPU engines are available immediately: flux analysis, annotation, codon work, primers, alignment, assembly, export. The catalog — 1,000 constructs across 65 organisms, 14.1 Mbp of designed DNA — is browsable as reference material whether or not you ever design anything in it.
Credits are only needed for GPU work, bought as you need them rather than committed up front. Groups that want API access, webhooks, reserved GPU capacity or SSO can move to an enterprise workspace, though most labs never need to.
Ready to run it
Try this in the Splice Platform.
CPU engines are free to run. GPU engines run on prepaid credits.
Related