Self-hosting gives you full control over where your data lives and how the platform is deployed, while running the same Tines 3B that powers our cloud.

If you're considering deploying Tines 3B in a self-hosted environment, please first contact your account executive or customer success manager to explore your options.

Following that conversation, you'll be put in touch with the right technical resources to grant you access to the Tines 3B images, guide you through the installation, and provide ongoing support.

Tines 3B self-hosted artifacts are delivered as a set of OCI-compatible container images from our registry at oci.tines.com. The registry supports standard tooling such as Docker, Podman, containerd, Kubernetes, and ORAS, along with version tags, digest-based references, and Cosign signature verification.

For air-gapped environments, we also provide a self-contained offline bundle, so the servers running Tines 3B never need registry access.

There are two ways to run Tines 3B self-hosted:

Single server: the entire Tines 3B stack runs on one Linux virtual machine. This is the fastest way to get up and running, and it's what we recommend for evaluations, proofs of concept, and pilots. Because everything runs on a single box, there's no high availability or redundancy: if the server goes down, so does 3B.

Kubernetes: for production, we recommend deploying Tines 3B to a Kubernetes cluster using our Helm chart. This gives you high availability, horizontal scaling, and rolling upgrades, and fits naturally into existing cluster operations, monitoring, and access controls.

A common path is to start with a single-server install for the proof of concept, then move to Kubernetes for production.

Tines 3B requires a customer supplied large language model to function.

Connect the provider of your choice with your own API keys, such as OpenAI or Anthropic, use a cloud service in your own account, such as AWS Bedrock, or run a model within your own infrastructure. Tines 3B works with the model you choose, wherever it runs.

This means you choose the model, and your prompts and data flow directly between your Tines 3B deployment and your model provider, under your own agreement with that provider, never through us.

Tines 3B is delivered as standard container images, and we support deploying it on the following platforms:

For evaluations and proofs of concept, the single-server installer sets up everything needed on one Linux virtual machine, including the container runtime.

Other container orchestration systems that satisfy our reference architecture and sizing requirements may work, but the correct operation of that infrastructure remains the customer's responsibility.