GO LOCAL.AI
Menu

Your data.
Your boundaries.

Deploy the Go Local AI Platform on your infrastructure, in private cloud or with managed hosting. Set where workloads can run and how their data can be used.

Private deployment and approved external routes Routing policy keeps private workloads inside the customer-controlled deployment. The gateway can also call a cloud API when that route is approved. The same platform can run on-premise, in private cloud or on managed infrastructure. YOUR DATA BOUNDARY Your gateway + Go Local AI Workload classification + routing policy Private small Routine work Private large Harder work Cloud / API Approved route EXTERNAL ON-PREMISE · PRIVATE CLOUD · MANAGED Private deployment and approved external routes Routing policy keeps private workloads inside the customer-controlled deployment. The gateway can also call a cloud API when that route is approved. The same platform can run on-premise, in private cloud or on managed infrastructure. YOUR DATA BOUNDARY Gateway + Go Local AI Classification + routing policy Private small Routine work Private large Harder work Cloud / API External, approved routes ON-PREMISE · PRIVATE CLOUD · MANAGED

Set the boundary
before data moves.

Before Go Local AI receives client data, we sign an NDA and data processing agreement. For managed deployments, processing locations and subprocessors are named before data moves.

Location is explicit.
Our standard managed deployment uses Civo’s UK facilities. When you run the software on your own infrastructure or private cloud, you choose the location.
Cloud routes need approval.
An external API route is optional. Its provider, processing country, retention and training terms are agreed before use. UK-only processing requires every approved route to stay in the UK.
Training your routing classifier.
All data used to train and evaluate your routing classifier stays within your data boundaries. Selected workload examples are anonymised, cleaned and labelled there. Client data is never used to train a model for another client.

Shared or dedicated.

Logical isolation.

Managed deployments can share hardware and inference weights. Each client has separate API credentials, logs, request context and retrieval. Access is enforced at the gateway.

Physical isolation.

Dedicated hardware for one client, with the agreed gateway controls. Used where the workload or data requirements call for physical separation.

What data is kept and who can access it.

Platform data stays within your deployment. Running the software does not give Go Local AI access to your prompts, responses or documents.

Inference.
Prompts and responses are processed by the models in your deployment, or by external APIs you have approved. Temporary model memory is part of that processing.
Monitoring and training.
Usage records and selected workload examples may be retained within your data boundaries for monitoring, evaluation and classifier training. What is retained and for how long is agreed for your deployment.
Managed hosting.
Data is processed within your hosted environment. The operational logs visible to Go Local AI are anonymised.
Assessments.
For an assessment, you share the agreed usage data and work samples with us. Access and retention are agreed before transfer and covered by the NDA and data processing agreement.

How the platform
is deployed.

Deployed by us.
We install and configure the platform on your infrastructure. Your workload data stays within your environment and is not visible to us.
Deployed by you.
Your team installs and runs the engines on your infrastructure or private cloud. You control the platform, its access settings and its data.
Managed.
We host and operate the platform and any agreed private models. Client data stays within the hosted data boundaries; we see only anonymised operational logs.
Changes.
Infrastructure changes require approval. Automatic classifier deployment can be enabled after evaluation and shadow testing; otherwise, each update requires approval.
Ask about data processing Website privacy notice