Safety needs evidence people can inspect.
We treat AI safety as a public-good responsibility. Systems that act in the world need boundaries, appeal paths, accountable records, and runtime checks that operators can inspect before something goes wrong.
Creed Space turns those commitments into tools people can run, inspect, challenge, and share: local evaluation, signed policy context, fleet audit trails, and public research with clear caveats.
AI capability is here. Trust infrastructure isn't.
That's the bottleneck. That's what we build.
The technology works. The question is whether it can be trusted.
Governance infrastructure must keep pace with capability: the trust layer that turns impressive demonstrations into reliable, deployed systems.
Governance at human speed cannot govern technology at AI speed.
AI capabilities improve in months. The governance layer itself must be AI-augmented.
You can't govern a system faster than you by standing outside it.
Governance has to run where decisions happen. Human values set the direction, and automated checks help enforce them at scale.
Personalised constitutional AI is to the AI era what SSL was to e-commerce.
E-commerce took off because trust infrastructure made it trustworthy. AI deployment requires the same.
Compliance is retrospective. Governance is prospective.
Constitutional constraints enforced at the moment of inference, not reviewed in a quarterly report.
How we treat AI now establishes patterns for everything that follows.
Good infrastructure makes careful defaults easier than unsafe shortcuts.
What operators ask us most
Is there a cost?
Guardian local tools are available under the Apache 2.0 license. Hosted services can require accounts, provider credentials, usage limits, or charges set for the active deployment.
Is Creed Space an enterprise software company?
Creed Space pursues a public-good mission. Its legal operator and organisational status remain pending in the current Terms and Privacy publication candidates.
Are there limits on agents or usage?
The published local Guardian code has no licence-metered evaluation limit. Capacity and provider limits depend on your hardware and configured services.
What have you open-sourced?
Guardian is Apache 2.0 licensed and VCP is an open standard. Each published repository and research artifact states its own licence and release status.
How is this sustainable?
Local Guardian deployments use the operator's hardware and selected model. Hosted operation has infrastructure and people costs; active pricing and funding are deployment-specific.
Does this support EU AI Act compliance?
Fleet source and design support map risk management (Article 9), record-keeping (Article 12), human oversight (Article 14), and robustness monitoring (Article 15). These mappings are technical aids. Exact deployment coverage, provider retention, legal interpretation, and independent assurance remain open.
How can my organisation get involved?
Deploy Guardian across your AI fleet. Integrate VCP into your products. Sponsor development. Fund safety research. Contact us to explore partnership.
For AI providers
The published VCP specification can be implemented without using a Creed Space hosted service. Any hosted support, provider, or infrastructure terms are separate.
Learn more