The kernel is free. Forever.
Run VAIXLNS yourself at no cost. Pay only when you want us to run the control plane, keep the ledger, and stand behind it.
For builders, researchers and anyone who wants to read every line before trusting it.
- Full kernel: canonical events, state, transition, replay
- Intelligence Fabric and VX subproject router
- Ratified specifications and the 32-week plan
- GitHub issues and pull requests
- Unlimited self-hosted nodes
For teams putting agents into production who need the ledger kept, the roots verified, and the constitution reviewed by someone else.
- Everything in Open Source
- Hosted control plane and managed event ledger
- Replay verification service with signed state roots
- Governance console: constitutions, actors, authorities
- SSO, audit export, retention policies
- Email support with a two-business-day response
For regulated and sovereign environments where the runtime must prove itself to an auditor, not just to you.
- Everything in Team
- On-premises or air-gapped deployment
- Custom kernels and constitution authoring
- Compliance evidence packs from the ledger
- Dedicated engineer and architecture reviews
- Priority roadmap input
What each plan includes.
| Capability | Open Source | Team | Enterprise |
|---|---|---|---|
| Deterministic kernel and replay engine | yes | yes | yes |
| Intelligence Fabric (DAG capability runtime) | yes | yes | yes |
| Self-hosted nodes | unlimited | unlimited | unlimited |
| Hosted control plane | — | planned | planned |
| Managed, append-only ledger with signed roots | — | planned | planned |
| Governance console | — | planned | planned |
| On-premises / air-gapped | DIY | — | planned |
| Compliance evidence packs | — | — | planned |
| Support | community | dedicated engineer |
Straight answers.
Is the open source version limited in any way?
No. Everything in the repository is the real kernel. Paid plans add services around it: hosting, verification, governance tooling and support. The code that proves, ledgers and replays is the same in every plan.
When do Team and Enterprise become available?
They are planned, not shipped. We are building the hosted control plane on top of the alpha kernel and will open Team to the waitlist first. Joining the waitlist costs nothing and commits you to nothing.
Can we run VAIXLNS entirely inside our own network?
Yes. The kernel has no external dependencies beyond Python and Node, and never calls out. Air-gapped deployment with support is part of the Enterprise plan.
Which AI models does it work with?
VAIXLNS is model-agnostic by design. The Intelligence Fabric composes capabilities as a DAG, and a capability can wrap any model or API. The governed runtime sits around the model, so the proof-before-effect rules apply regardless of which model proposes an action.
Who builds VAIXLNS?
Team Pro Workload. The project is developed in the open on GitHub; issues and pull requests are welcome.