OpenAI Compatibility

THROUGHPUTS is wire-compatible with the OpenAI Chat Completions and Embeddings APIs. Any client written for OpenAI — official SDKs, community libraries, LangChain, LlamaIndex, Continue, Aider — works with THROUGHPUTS by changing one line.

The one-line migration

- base_url="https://api.openai.com/v1"
+ base_url="https://api.throughputs.dev/v1"

That's it. Same request shape, same response shape, same streaming format, same error codes.

Python

from openai import OpenAI

client = OpenAI(
    api_key="thp_live_xxx",
    base_url="https://api.throughputs.dev/v1",
)

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello"}],
)

TypeScript

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.THROUGHPUTS_API_KEY!,
  baseURL: "https://api.throughputs.dev/v1",
});

const response = await client.chat.completions.create({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Hello" }],
});

LangChain

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="gpt-4o",
    api_key="thp_live_xxx",
    base_url="https://api.throughputs.dev/v1",
)

What works unchanged

• Chat completions (streaming and batch)
• Function / tool calling
• JSON mode (response_format)
• Vision (multimodal image_url content parts)
• Embeddings (including adjustable dimensions)
• Rate limit semantics (429 + Retry-After)
• Error shape ({"error": {"message", "type", "code"}})

What's different

Model names

THROUGHPUTS accepts both the provider's native slug and normalized slugs:

Aliases are stable — THROUGHPUTS resolves them to the latest version of the underlying model. Use full slugs if you need to pin a specific version.