Develop your process on a model. Confirm it in the lab.
Make the right process decisions faster, with up to 70% fewer experiments. Novasign Studio guides experimental design, connects your data and builds hybrid models across your entire process chain, from upstream and filtration to chromatography. Predict how changes affect titer, purity and yield before running experiments, then use traceable results to guide scale-up and support regulatory submissions, all in one platform.
For biopharma, CDMOs and industrial biotech, including precision fermentation and cultured food.
Reactor · TFF · Chromatography
Approve for use
Reactor Model, trained on the new data
ApproveFrom broken handovers to one seamless, flexible workflow
Every bioprocess project runs through workflows that cross tools, units and teams: from the experimental plan to a decision, from upstream to downstream, and from early development to manufacturing. Each one breaks at its handovers. Novasign Studio turns them into one workflow that you adapt to how your team works.
Planning, execution, data, modelling and decisions usually live in separate tools: a planning tool, the control system, spreadsheets or databases to aggregate and inspect the data, a modelling tool and a meeting for the decision. Each tool needs its own expertise, and every handover is done by hand.
- Conventionally, experimental planning reaches the experiments through manual setup.
- The results reach spreadsheets and databases by copy and paste.
- The data reaches the models after being aligned by hand.
- The decision that follows is hard to retrace.
How Novasign Studio clears the way
From plan to execution
Conventionally a design is set up on the equipment by hand, sometimes with the help of DoE software. Together with device vendors, Novasign is building the link that sends a design from Novasign Studio straight to the equipment, without a manual method setup.
From experiments to data
In Novasign Studio every step is a node on one canvas. Data importer nodes read files and live process data, and curation nodes merge them into one dataset. In a biosimilar project with Repligen, Novasign Studio combined bioreactor data, offline viable cell density and titer, MAVEN glucose control and ZipChip CE-MS quality data in one workflow.
From data to model
Extensive preprocessing nodes filter, align and validate the data. You build a reliable ETL pipeline for your process development once, and it runs the same way for every campaign that follows.
From model to decision
The model behind the digital twin simulates the scenarios you have not run in the lab and ranks them by the objectives you choose, several at once. Novasign Studio is collaborative, so the whole team opens the same model, works on it together and can retrace the decision.
From decision to the next experiments
Model-based DoE proposes the next experiments from the digital twin’s simulations. Tell it how many runs the next round holds, such as 4, 8 or 12 vessels of a parallel bioreactor system, and it recommends that many. Across customer programs this has meant 35–70% fewer experiments.
Upstream and downstream are often developed by separate teams, each with its own responsibilities, background knowledge and vocabulary. A change in the reactor reaches the next unit as a surprise, because nobody sees which outputs of one unit drive the next.
- Conventionally, the bioreactor, centrifugation, chromatography, TFF and formulation are developed as separate steps.
- Each handover goes to a separate team with its own vocabulary, and no model is shared.
How Novasign Studio clears the way
A model for every unit operation
The bioreactor, TFF and chromatography have prebuilt models in Novasign Studio. Centrifugation and formulation are modelled with your own model today, and prebuilt models for both are in development.
Units linked by the model and the digital twin
The models of neighboring units are connected, so the digital twin shows which outputs of one unit affect the next and the whole process is understood in one environment. In the ECONTI consortium, the twin of a continuous process predicted the load reaching the chromatography and set the switching of its columns.
Between development and manufacturing the separation is often intended: every stage owns its responsibilities and its expertise. The model built in development rarely reaches the next stage, so each stage starts again from its own data.
- Conventionally, early development, late development, pilot and scale-up, and manufacturing each start again from their own data.
- Scale-up meets a new scale unprepared, and the technology transfer hands on documents rather than a model.
How Novasign Studio clears the way
One model that grows with the project
Workflows in Novasign Studio are reusable and extendable, so the model from early development is retrained with the data of each later stage instead of being rebuilt. The process knowledge stays in decision workflows that carry over to the next stage and team.
Predictive scale-up instead of reactive troubleshooting
For the bioreactor, the model works together with computational fluid dynamics from Novasign’s technology partner SimVantage to predict process performance and product quality at scale.
Backed by customer results and peer-reviewed science
Bespark*bio
Experiment reduction and optimization
A model trained on 8 runs instead of 22 made the same decisions for optimizing a viral-vector process as one trained on the full campaign.
64% fewer experiments, about €40,000 of raw material and 12 weeks saved.
Bisy
Industrial process optimization
Two rounds of designed experiments optimized a Komagataella phaffii process for an industrial enzyme.
Enzymatic activity increased 2.4-fold.
Arkeon
Optimization and scale-up
A designed campaign, model-proposed conditions and validation runs, followed by a scale-up prediction.
5.3-fold specific yield at 150 L over the initial process, and a successful scale-up.
Repligen
Matching CQAs for biosimilars
A biosimilar process was designed for a target glycan pattern, with Repligen’s process analytics integrated.
A design space for glycoform and glycation, mapped from five experiments and a hybrid model.
ECONTI consortium
E2E process
Hybrid models of four unit operations ran inside the automation platform of a continuous E. coli process and set the chromatography switching of the next cycle.
Intensified DoE raised the upstream titer by 60% over the original process conditions, and the process ran end to end with model-based control.
From the beginning, bespark*bio has worked closely with Novasign… This transition is powered by Novasign Studio, which we use as our exclusive modeling environment.
Why teams choose Novasign
The next round sized to your vessels
Experiment designA method published in peer-reviewed journals
PublicationsNamed customer results
Case studiesA team that can run the experiments for you
Process development
Novasign was founded in August 2019 as a BOKU University spin-off, based in Vienna, Austria. It makes Novasign Studio, browser-based bioprocess modelling software.
Start with an idea, and let the AI Assistant build it with you
Describe what you want to find out, and the optional AI Assistant sets up the workflow with you, from your data to a digital twin. You reach a first model faster, without knowing every node, and every suggestion stays yours to check and accept.
Built for your industry
- CDMOsDeliver faster, well-justified process decisionsTurn process data, results and expertise into workflows your team reuses for every client.
- Fewer experiments per client program
- Workflows that carry over to the next project
- Decisions your sponsors can follow step by step
- BiopharmaStop running experiments you have already runMake the knowledge from earlier campaigns and platform molecules usable for the next decision.
- Reuse historical data across molecules and programs
- Bring your existing models, scripts and code
- Upstream and downstream in one connected workflow
- Industrial biotechReach viable unit economics fasterFind the process variables that move the cost per gram, and scale up with confidence.
- See which conditions drive production cost
- Recommendations your team can inspect and challenge
- Scale-up predictions before the expensive run
Why teams like working in Novasign Studio
One workflow, from the data to the model, built in the browser and shared by the whole team.
Work together
Colleagues edit the same workflow at the same time, and a node someone is working on is locked for everyone else.
Nothing to install
Collaborators open the project in their browser and look at the same data, model and result.
Every result traceable
Every change carries its author and time, and every training run is recorded with its data and parameters.
No code, or your own code
Build a workflow by dragging nodes, or write your own Python in the same workflow.
Hybrid models learn from less data and respect the mechanistics
A hybrid model combines a given mechanistic structure (mass balances, stoichiometry) with a neural network that learns only the reaction kinetics. Because the mechanistics you know are already in the model, it has less left to learn, so the data from an affordable campaign is enough to train it, and it can often predict beyond the conditions it was trained on.
- Inside the Mass Balance. c t+1 V t+1 = c t V t + Σ c in V in − c t Σ V out
- Process Parameters. Temperature, pH
- Feed. Glucose
- Samples. Volume removed
- Mass Balance. Given · mechanistic
- Machine Learning. Learned · from data. Rates of change
- Kinetics. Given · mechanistic. New states
- Predicted. VCD. Titer. Glucose
Every experiment improves the model
Each experiment adds to the model. Novasign Studio proposes the runs to perform, the scientist can adapt them, and the results are used to retrain the model. Fewer experiments are needed because the knowledge of the last campaign is carried over.
Novasign Studio proposes, the scientist adapts
The model proposes the runs to perform and shows how confident it is. Your team can adapt them before they go to the bioreactor.
8 runs instead of 22
to optimize a viral-vector process (Bespark*bio)
- State the goal, for example the titer, a quality attribute or the operating window the campaign has to reach.
- Design the experiments from what is already known.
- Run the experiments.
- Prepare the data. Alignment, units and derived quantities are nodes in the workflow, not a script on one person’s laptop.
- Train the model on the campaign. A person starts every training run.
- Simulate the runs that were not carried out and read a candidate optimum from the result. A scientist decides which experiments to run next, their results join the dataset, and the model is retrained. Later rounds often need only a confirming run rather than another campaign.
Start with one process question. See what your data can do.
The right starting point if you want concrete results. We work with you on a scoped proof of concept using your own process data, so the outcome is relevant to your process from the first step.
- 1
Scoping call
We sign an NDA with you. Then you define one process challenge and one unit operation or workflow to focus on.
- 2
Data review
Novasign reviews your existing data and defines what the proof of concept can produce, at no cost up to this point.
- 3
Proof of concept
We build and run the workflow in Novasign Studio together with you.
- 4
Results
Clear outputs, a defined scope, and your decision on how to proceed.
- 5
Handover
The project is handed over to you in Novasign Studio, so your team continues from where the proof of concept ended.
Studio runs where your data is
Novasign Studio runs on your own servers, in your own cloud account or on an appliance that Novasign builds and ships, and the on-premise and appliance options can run air-gapped. You decide where your process data is stored and what leaves your network.
- On-premise
- Your own data centreavailable today
- Your cloud
- Your own cloud accountavailable today
Services that get you to results faster
From experiments with our partners to a proof of concept in Novasign Studio and training for your team, our services meet you where your project stands.
Process development
No capacity for experiments, or no data at hand? With our partners in Vienna we carry out the experimental work, planned with you so the data is fit for modelling.
Process developmentProof-of-concept modelling
Our engineers build the first model of your process with you in Novasign Studio, so your evaluation is faster and the results are yours to continue from.
Proof-of-concept modellingNovasign Studio training
Remote or on site, on your own data after a proof of concept or on Novasign’s upstream and downstream datasets without one.
Novasign Studio trainingProcess modelling workshop
A hands-on course at our offices in Vienna, on real data from upstream through filtration to downstream. The next one is planned for the end of Q1 2027.
Process modelling workshopScientific modelling
A new or intensified process, an unusual combination of unit operations, or an idea you want to evaluate before you build it. Our bioprocess engineers design and model the unit with you.
Scientific modelling
From the blog
Articles on methods and results, written by the people who build the models.
BioprocessBioprocess DevelopmentMar 20, 2026De‑Risking Bioprocess Scale‑Up: Predictive Modeling for Reliable Fermentation and Cell Culture Manufacturing
Scaling a bioprocess from laboratory development to industrial manufacturing is one of the most critical, and risky steps in biotechnology. For contract development and manufacturing organizations (CDMOs), large pharmaceutical companies, industrial biotechnology producers, and emerging sectors such…
BioprocessBioprocess DevelopmentMar 20, 2026Autonomous Continuous Biomanufacturing: Closing the Gap Between Process Development and Production
Continuous biomanufacturing has long been considered one of the most promising approaches to increase productivity, reduce facility footprint, and enable more sustainable bioprocessing.
BioprocessBioprocess DevelopmentMar 20, 2026Next‑Generation Media Development: How Modeling Reduces Cost and Accelerates Bioprocess Optimization
Media composition is one of the least transparent elements in modern cell culture bioprocess development. Whether producing monoclonal antibodies, recombinant proteins, industrial enzymes, or cultivated food products, the composition of the cell culture or fermentation media directly determines…
See how Novasign fits your process and your data
30 minutes on your own campaign data, no qualification step.