@standardagents/deepinfra
v0.40.6
Published
DeepInfra provider for Standard Agents
Readme
@standardagents/deepinfra
First-party DeepInfra provider for Standard Agents.
pnpm add @standardagents/deepinfraimport { defineModel } from '@standardagents/builder';
import { deepinfra } from '@standardagents/deepinfra';
export default defineModel({
name: 'deepinfra-chat',
provider: deepinfra,
model: 'deepseek-ai/DeepSeek-V3',
});Set DEEPINFRA_API_KEY for direct BYOK requests. The provider uses DeepInfra's OpenAI-compatible Chat Completions endpoint, discovers chat models and per-million-token prices from the live public catalog, supports streaming, images, tools, strict structured output, reasoning controls, prompt caching options, and request inspection, and preserves DeepInfra's response-level usage.estimated_cost as usage.cost.
Model discovery also resolves the model namespace (for example deepseek-ai, meta-llama, or Qwen) to bundled local lab artwork, so Builder model lists show the model creator rather than repeating the DeepInfra provider mark.
Connection slots are apiKey, baseUrl (default https://api.deepinfra.com/v1/openai), modelsUrl (default https://api.deepinfra.com/v1/models), and timeout.
See the DeepInfra provider guide and DeepInfra API docs.
