Connect every artifact. Ground every decision in integrated technical truth.
The systems engineering stack that teams use to design, maintain and verify complex products. Requirements, architecture, parameters and verification runs stay linked as the design evolves. AI-native. Understanding the impact of a change takes seconds, not weeks.
- ARC available now
- More coming soon
Ontological Systems Engineering.
Applied.
Your ontology defines your engineering vocabulary: what exists, what it contains, how it connects and which rules apply. Dymension makes that ontology the runtime language of your system model, connecting data from other tools and powering the processes built on it.
All the systems engineering artifacts share this language. Humans and AI agents work from the same connected context, even as your processes evolve.
The next generation of engineering intelligence runs on connected context.
Everything ARC does today, on one connected model.
ARC is our first product: the workspace for data-driven systems engineering.
Requirements as data
Write and manage requirements that hold their links to everything they touch. When one changes, the things that depend on it know. Owners, responsibilities, and deadlines live right on the record, fitted to your processes.
For the teams building the hardest things, software shall be the operating system of their processes.
We build software that keeps engineers on the engineering and decision-makers on the latest state of development. Each discipline works on shared system data, on its own terms. Humans and AI need that shared foundation to engineer together, and Dymension is it: one knowledge graph of the engineering data, with the decisions behind it.
- 01
The home for system context.
Dymension centralises the engineering data your programme depends on. Disciplines keep their tools and meet the data on their own terms.
- 02
Systems engineering, finally with an operating system.
Requirements, architecture, interfaces and verification share one foundation, so the impact of a change is a question you ask, not a week you spend.
- 03
Nothing gets lost. Across the lifecycle.
Every version, decision and result is kept and linked, from concept through design to test. What was created and proved in one programme is reused by the next.
- 04
AI that understands your system.
Agents work on the structure of your system rather than on scattered files, and every answer traces back to the model.
For problems that do not fit in a spreadsheet, from one team to an entire supply chain.
Made for programmes that cannot afford a leak or a lock-in.
EU hosting by default
Cloud plans are hosted in the EU, and the optional AI assistant uses European models by default. Other models and deployment options are available.
Your deployment, your rules
Cloud, private cloud, on-premise, sovereign or air-gapped, for programmes under export control or classification.
Access follows your structure
Permissions follow your own project structure, so each person sees the part of the model they are entitled to.
Open by design
The model follows SysML v2 semantics. Export in SysML v2 notation, and to OWL and SHACL, is on the way, so your model is never stranded in our representation of it.
ARC is available now.
Free
For evaluating ARC and for small teams getting a first project connected.
Request access- Up to 10 people
- 3 viewer seats, free
- EU cloud
- Community support
- No AI assistant or document extraction
- No DymSDK or API access
- Lower upload and import limits
Pro
For teams running programmes on ARC, with AI and the whole feature set.
Request pricing- Uncapped seats
- Uncapped viewer seats, free
- EU cloud
- AI assistant and document extraction, with credits included per seat
- DymSDK and API access
- Higher upload and import limits
- Priority support
Enterprise
For organisations where the deployment is the hard part, on your own infrastructure.
Contact us- Everything in Pro
- Cloud, private cloud, on-premise, sovereign or air-gapped
- Uncapped document import
- Upload limits raised per organisation
- SSO and SAML
- SLA and dedicated support
- AI runs on cloud deployments only
The questions engineering teams ask us first.
ARC is a systems engineering workspace where requirements, architecture, verification and decisions live in one connected model instead of in separate documents, so a change in one place is visible everywhere it matters.
Requirements tools manage one artifact type well and hand everything else to exports and spreadsheets. Dymension is not only that, it models the relationships themselves: the link between a requirement, the component that satisfies it, the test that verifies it and the decision that shaped it. Traceability becomes a property of the data, not a report somebody maintains by hand.
No. Teams usually start with one subsystem or one programme and connect the rest as they go, so the model earns its place before it becomes the system of record. Usually the first step is to start with an ontology to manage the process that brings more issues and inconsistencies to the team.
Documents store text; a graph stores meaning. Because every requirement, component and decision is a node with typed relationships, "what breaks if this changes" is computed from the model rather than reconstructed by a person, and AI agents can reason over that structure instead of guessing from prose.
ARC follows SysML v2 semantics, so the model you build here is not stranded in our representation of it. Export in SysML v2 notation, and a SysML v2 compliant API, are both on the way, so the model can leave at any point in a form other tools already understand.
The model is not stored in OWL, ORM or SHACL. Its semantics are deliberately close to them, and export to OWL or ORM, with the constraints that keep it valid coming out as SHACL, is on the way. The point is portability: the meaning of your model, and the rules it has to satisfy, should not be locked to our representation of them.
Access follows your own project structure, so each person sees the part of the model they are entitled to. For programmes with export-control or classification constraints we also offer on-premise, sovereign and air-gapped deployments.
See the deployment optionsA short scoping call, then a pilot on a real subsystem, so you judge ARC on your own data rather than on a demo.
Something not answered here?Talk to us