Onshore

Inference and data stay in Singapore.

Latest models

Current releases, live now.

No premium

Vendor pricing, hosted locally.

Singapore-hosted AI models at the same cost.

You pay what the model vendor charges. The difference is where your data goes.

DeepSeekSG-hosted

V4.1 Flash

Fast, low-cost general model

  • Vendor direct$0.30 / $1.20, offshore
  • AWS BedrockNot offered onshore
$0.30 input
$1.20 output
Z.aiSG-hosted

GLM-5.3

Frontier-level reasoning and coding

  • Vendor direct$1.40 / $4.40, offshore
  • AWS BedrockNot offered onshore
$1.40 input
$4.40 output

USD per 1M tokens, input / output, as at September 2026.

What you can build.

Search and RAG
Customer assistants
Agents and tool use
Coding assistants
Summarisation
Image understanding

Switch in two lines.

Your existing OpenAI code works as is. Change the base URL and the key, keep everything else.

from openai import OpenAI

client = OpenAI(
    base_url="https://cloud.tuas.ai/v1",
    api_key="TUAS_AI_API_KEY",
)
# requests stay in Singapore

More models on request.

Any open model, or dedicated inference for your workload.

QwenKimiGemmaGLMAnd more

Sovereign by default.

Runs in Singapore

Models and inference stay onshore.

No training

Your data never trains any model.

No retention

Requests are processed, then discarded.