Inference overview
Galliun Inference is a paid, OpenAI-compatible AI API. You pay per request in USDC from a connected wallet.
What you can do
- Chat completions (
client.inference.chat) - List models and pricing catalog
- Multimodal (vision) chat
- Tool definitions on supported models
- Managed conversations (wallet auth required)
- Embeddings (
client.inference.embeddings.create)
Setup
Pass an inferenceWallet when creating the client. See Create a client and Connect a wallet.
const response = await client.inference.chat({
model: "openai:gpt-4o-mini",
messages: [{ role: "user", content: "Hello" }],
});
Model IDs
Use provider:model format, for example openai:gpt-4o-mini, anthropic:claude-sonnet-4-20250514. Providers include openai, anthropic, deepseek, and qwen. List live IDs with client.inference.models.list().
Payments
Inference uses the exact scheme: one USDC transfer to the Galliun treasury. The SDK handles the 402 → pay → retry flow. See Core concepts.