Usage & examples.

Non-streaming request:

curl https://uai.sh/v1/chat/completions \
  -H "Authorization: Bearer uai-YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4.1-mini",
    "messages": [{"role": "user", "content": "Say hi in one short sentence."}],
    "stream": false
  }'

Streaming (SSE):

curl https://uai.sh/v1/chat/completions \
  -H "Authorization: Bearer uai-YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4.1-mini",
    "messages": [{"role": "user", "content": "Say hi in one short sentence."}],
    "stream": true
  }'

List available models:

curl https://uai.sh/v1/models \
  -H "Authorization: Bearer uai-YOUR_API_KEY"

Install the OpenAI SDK:

pip install openai

Basic request:

from openai import OpenAI

client = OpenAI(
    base_url="https://uai.sh/v1",
    api_key="uai-YOUR_API_KEY"
)

response = client.chat.completions.create(
    model="openai/gpt-4.1-mini",
    messages=[{"role": "user", "content": "Say hi in one short sentence."}],
    stream=False
)

print(response.choices[0].message.content)

Streaming:

stream = client.chat.completions.create(
    model="anthropic/claude-3.5-sonnet",
    messages=[{"role": "user", "content": "Write a poem"}],
    stream=True
)

for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="", flush=True)

Environment variable (recommended):

import os
from openai import OpenAI

# export OPENAI_API_KEY=uai-YOUR_API_KEY
# export OPENAI_BASE_URL=https://uai.sh/v1
client = OpenAI()  # reads env vars automatically

Install the OpenAI SDK:

npm install openai

Basic request:

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://uai.sh/v1",
  apiKey: process.env.UAI_KEY,
});

const res = await client.chat.completions.create({
  model: "openai/gpt-4.1-mini",
  messages: [{ role: "user", content: "Say hi in one short sentence." }],
  stream: false,
});

console.log(res.choices[0].message.content);

Streaming:

const stream = await client.chat.completions.create({
  model: "anthropic/claude-3.7-sonnet",
  messages: [{ role: "user", content: "Write a poem" }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content || "");
}

curl (free, nomic-embed-text via Mac Ultra):

curl https://uai.sh/v1/embeddings \
  -H "Authorization: Bearer uai-YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "shared-local/up_mac_ultra_local_ollama/nomic-embed-text",
    "input": "The quick brown fox"
  }'

Python (OpenAI SDK):

from openai import OpenAI

client = OpenAI(base_url="https://uai.sh/v1", api_key="uai-YOUR_API_KEY")

resp = client.embeddings.create(
    model="shared-local/up_mac_ultra_local_ollama/nomic-embed-text",
    input=["Hello world", "Another sentence"]
)
vectors = [d.embedding for d in resp.data]

Node.js (OpenAI SDK):

const resp = await client.embeddings.create({
  model: "shared-local/up_mac_ultra_local_ollama/nomic-embed-text",
  input: ["Hello world", "Another sentence"],
});
const vectors = resp.data.map(d => d.embedding);

Token usage in the final chunk — many SDKs (LangChain, OpenRouter SDK, llama_index) need this:

{
  "model": "openai/gpt-4.1-mini",
  "messages": [{"role": "user", "content": "hi"}],
  "stream": true,
  "stream_options": {"include_usage": true}
}

Last non-DONE chunk will include a populated usage object with prompt/completion/total tokens.

Cancellation — abort the request, the worker stops generating within ~1 s and releases the balance hold:

const ac = new AbortController();
setTimeout(() => ac.abort(), 200);
await client.chat.completions.create(
  { model: "openai/gpt-4.1-mini", messages: [...], stream: true },
  { signal: ac.signal }
);

finish_reason arrives in the last non-DONE chunk: stop | length | tool_calls | content_filter.