# Systellar Space > Systellar Space is a Barcelona company that builds synapse™, the operating system for fast-moving hardware teams. synapse™ is a model-based systems engineering (MBSE) platform built on SysML v2, one live model where engineers and AI agents design, verify and deliver complex systems. Requirements, architecture, verification, workflows and version control live on one system model that engineers edit in trees and tables, without writing SysML, and that AI agents can read and act on under human review. Teams adopt synapse™ in days, not months. ## Product - [synapse™ overview](https://systellar-space.com/): What synapse™ is and the hardware teams it is built for. - [Architecture](https://systellar-space.com/product/architecture): Model parts, interfaces, parameters and behaviour as one SysML v2 system model, in trees, tables and diagrams, without writing SysML. - [Requirements](https://systellar-space.com/product/requirements): Requirements as part of the system model: design constraints checked against changes, a process configured in minutes and traceability from mission need to component. - [Verification](https://systellar-space.com/product/verification): Verification campaign planning and tracking: actions linked to the requirements they prove, planned against milestones, with evidence attached and status that rolls up. - [Workflows](https://systellar-space.com/product/workflows): Automation of routine engineering work on the system model: simulations and budgets in Python, scheduled status reports and document generation, on demand, on a schedule or when an attribute changes. - [Version control](https://systellar-space.com/product/version-control): Version control for a system model: every change is made on a branch and merged through a reviewed merge request, with field-level diffs, tagged versions that can be compared and reusable libraries. - [AI Agents](https://systellar-space.com/product/agents): AI agents for systems engineering: an in-app agent edits the SysML v2 model within your permissions, and MCP opens the same project to external AI clients. - [Security and deployment](https://systellar-space.com/security): Self-hostable services, policy-based authorisation, sandboxed code execution and EU data and inference residency. ## Company - [Company](https://systellar-space.com/company): Who Systellar is and who synapse™ is for, plus the mission, values, founding team and the programmes and investors backing Systellar. - [Contact](https://systellar-space.com/contact): Book a synapse™ demo or talk to an engineer. ## Articles - [Your org chart is designing your product](https://systellar-space.com/blog/your-org-chart-is-designing-your-product): Conway's Law says your system mirrors your org chart. In a hardware startup, your team structure is quietly choosing your product architecture. - [Why hardware startups lose velocity as they grow](https://systellar-space.com/blog/why-hardware-startups-lose-velocity-as-they-grow): At some point, hiring more engineers stops making you faster. Most hardware startups hit this before they expect to, usually somewhere between the 20th and 40th hire. - [When your engineering process stops scaling](https://systellar-space.com/blog/when-your-engineering-process-stops-scaling): There is a team size where engineering stops working the way it used to. Most hardware teams hit it somewhere between 15 and 25 people. Few see it coming. - [Why 2-day MBSE adoption is not a feature: it is a business model](https://systellar-space.com/blog/why-2-day-mbse-adoption-is-not-a-feature): Every MBSE vendor talks about adoption. Most mean months of training, consultants, and rollout plans. Systellar means two days, and that difference is about who the tool is built to serve. - [Why AI cannot fix badly structured engineering data](https://systellar-space.com/blog/why-ai-cannot-fix-badly-structured-engineering-data): Every AI tool for engineering works well in the demo. The demo uses clean, curated data. Production engineering data is not clean or curated, and the AI does not know the difference. - [Why verification planning is painful: a structural explanation](https://systellar-space.com/blog/why-verification-planning-is-painful): Verification planning consistently takes longer than it should. The reason is not that writing test procedures is hard. Before any procedure can be written, someone has to reconstruct whether the requirements are still accurate, whether the design reflects them, and whether the test environment can address what is being asked. - [What a real digital thread requires in hardware engineering](https://systellar-space.com/blog/what-a-real-digital-thread-requires): In hardware engineering, the same critical parameter often lives in six different places and nobody knows which copies are current. A real digital thread is not about connecting tools. It is about one source that everything else reads from. - [Why most MBSE adoptions fail before they deliver value](https://systellar-space.com/blog/why-most-mbse-adoptions-fail): Most MBSE initiatives do not fail because the methodology is wrong. They fail because the tools required too much investment before delivering any value, and engineering teams stop investing when the return does not materialise. - [What real traceability requires: why linking artifacts is not enough](https://systellar-space.com/blog/what-real-traceability-requires): Most programs that describe themselves as having traceability are tracking links. A coverage matrix at 100% can still miss every requirements conflict that matters if the links are never checked for validity after the design moves. - [Why requirements management is not systems engineering](https://systellar-space.com/blog/why-requirements-management-is-not-systems-engineering): A requirements baseline can be perfectly managed and still fail to keep the design under control. The gap between tracking requirements and having them connected to the live system is where most programs lose ground. - [What version confusion actually costs an engineering team](https://systellar-space.com/blog/what-version-confusion-actually-costs): Version confusion rarely announces itself as a crisis. It accumulates quietly across shared drives and email attachments until a decision gets made from the wrong document at the worst possible moment. - [The real cost of manual change propagation in complex engineering](https://systellar-space.com/blog/the-real-cost-of-manual-change-propagation): When a parameter changes in a complex program, updating the number takes minutes. Tracing every artifact that depends on it can take days, and there is no reliable way to know when you are done. - [Why engineering coordination relies on people, not process: and why that breaks at scale](https://systellar-space.com/blog/why-engineering-coordination-relies-on-people): Most engineering programs don't fail because the engineers lack skill. They fail because the coordination infrastructure is built around people who know where everything is, and that breaks the moment it needs to scale. ## News - [Tumbleweed selects synapse™ to bring agentic hardware development to Earth-return capsules](https://systellar-space.com/news/tumbleweed-selects-synapse): Tumbleweed, the company making microgravity accessible for research and industry, engineers its Oasis spacecraft in synapse™ as one digital model that AI agents work on alongside its engineers. - [Hardware's Cursor moment: our talk at AI Summit Barcelona](https://systellar-space.com/news/hardwares-cursor-moment-ai-summit-barcelona-2026): Alexandre Gol i Mestre, Systellar's CTO, on why AI transformed software but not hardware yet: hardware has no codebase an agent can read. - [Systellar on AI for hardware at the AI Summit](https://systellar-space.com/news/systellar-at-ai-summit-barcelona-2026): Alexandre Gol i Mestre, Systellar's CTO, takes the Demo Stage on 22 September with "Satellites, Planes & Trains: Why Hardware's Cursor Moment Just Started". Why AI has barely touched hardware engineering, what has to exist for an agent to see a whole system, and why that exists now. - [Systellar wins the Santander X Award | Startup Universitaria 2026](https://systellar-space.com/news/systellar-wins-santander-x-award-2026): Systellar has been named a winner of the Santander X Award | Startup Universitaria 2026, recognised by Banco Santander's Santander X community in the university start-up category. Systellar is an active part of the entrepreneurial community around UPC and thanks Emprèn UPC and its partners for their initiatives. - [Systellar reaches the finals of the EU Startup Summit pitch competition](https://systellar-space.com/news/systellar-eu-startup-summit-2026-finals): 15 finalists selected from over 1,600 applications across Europe. Systellar pitched at the EU Startup Summit 2026 in Valletta, Malta alongside startups from aerospace, robotics, AI, and healthcare. - [Systellar reaches the London regional finals of The Pitch by Deel](https://systellar-space.com/news/systellar-the-pitch-london-finals): Systellar was selected to compete in the London regional finals of The Pitch by Deel, presented by JP Morgan. It is one of Europe's most recognised startup competitions for early-stage companies. - [Systellar at 4YFN](https://systellar-space.com/news/systellar-at-4yfn-2025): 4YFN is a great opportunity to connect with the entrepreneurial ecosystem and discover inspiring projects. Thanks to ESA BIC Barcelona for their support to startups like ours and for helping to drive the New Space ecosystem. - [Systellar joins Techstars Spring 2026 global cohort](https://systellar-space.com/news/systellar-joins-techstars-spring-2026): Systellar has been selected to join the Techstars London Spring 2026 Accelerator. The program begins with the global kickoff in New York City and will support growth of synapse™ and engagement with the Techstars network.