AI Inference Sovereign LLM API Billed in NZD
Your developers already know how to call an LLM API. AI Inference gives them the same interface they use today — drop-in OpenAI-compatible endpoints — but pointed at open-weight models running on ASI GPUs in New Zealand.
Put AI in your product without a US cloud contract
Drop-in compatibility, sovereign infrastructure
Built for teams who want to put AI into a product or an internal application without a US cloud contract, a US dollar invoice, or a privacy conversation they cannot win.
Change one base URL. Existing SDKs, frameworks and agent tooling work unmodified — with no data leaving the country and no prompts used to train anyone's model.
OpenAI-compatible • Prepaid NZD credits • Team and key-level controls • Full observability
Key Capabilities
Drop-in compatibility
Change one base URL. Existing SDKs, frameworks and agent tooling work unmodified.
Self-service portal
Sign up with SSO, generate and rotate API keys, and see spend in real time.
Transparent token metering
Separately metered input, cached input and output tokens — cached prompts cost less.
Prepaid credit model
Top up in NZD, consume against the balance. No surprise invoices, no exchange-rate drift.
Curated open-weight models
Current-generation instruct, reasoning, vision and embedding models, refreshed as the field moves.
Reserved capacity
Dedicated throughput for production workloads that cannot queue behind anyone else.
Packaging
Basic for prototyping · Enterprise for production
Basic — self-service signup, prepaid credits, shared capacity. Ideal for prototyping and internal tools.
Enterprise — contracted terms, invoiced billing, reserved capacity, private model deployments and support commitments.
Best for software teams, ISVs and digital agencies embedding AI into applications
Embed AI Without Sending Data Offshore
Best for any organisation whose privacy, procurement or regulatory position rules out sending prompts offshore.