One endpoint, five kinds of Arabic.

curl https://api.lahja.dev/v1/chat/completions \
  -H "Authorization: Bearer $LAHJA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemma-4-31b-it:arabic-gulf",
    "messages": [{"role": "user", "content": "وين أقرب صيدلية؟"}]
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.lahja.dev/v1",
    api_key=os.environ["LAHJA_API_KEY"],
)

r = client.chat.completions.create(
    model="google/gemma-4-31b-it:arabic-gulf",
    messages=[{"role": "user", "content": "وين أقرب صيدلية؟"}],
)
print(r.choices[0].message.content)
print(r.model)  # google/gemma-4-31b-it:arabic-gulf

// Node 18+. Run this on your server: the key must not reach a browser.
const r = await fetch("https://api.lahja.dev/v1/chat/completions", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.LAHJA_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "google/gemma-4-31b-it:arabic-gulf",
    messages: [{ role: "user", content: "وين أقرب صيدلية؟" }],
  }),
});
const j = await r.json();
console.log(j.choices[0].message.content);

// Or with the openai package:
// import OpenAI from "openai";
// const client = new OpenAI({ baseURL: "https://api.lahja.dev/v1", apiKey: process.env.LAHJA_API_KEY });

curl -N https://api.lahja.dev/v1/chat/completions \
  -H "Authorization: Bearer $LAHJA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "google/gemma-4-31b-it:arabic-egy",
       "stream": true,
       "stream_options": {"include_usage": true},
       "messages": [{"role": "user", "content": "إزاي أرجّع الأوردر؟"}]}'

stream = client.chat.completions.create(
    model="google/gemma-4-31b-it:arabic-egy",
    messages=[{"role": "user", "content": "إزاي أرجّع الأوردر؟"}],
    stream=True,
    stream_options={"include_usage": True},
)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)
    if chunk.usage:  # last chunk, empty choices
        print("\n", chunk.usage.prompt_tokens, chunk.usage.completion_tokens)

import OpenAI from "openai";
const client = new OpenAI({ baseURL: "https://api.lahja.dev/v1", apiKey: process.env.LAHJA_API_KEY });

const stream = await client.chat.completions.create({
  model: "google/gemma-4-31b-it:arabic-egy",
  messages: [{ role: "user", content: "إزاي أرجّع الأوردر؟" }],
  stream: true,
  stream_options: { include_usage: true },
});
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
  if (chunk.usage) console.log("\n", chunk.usage);
}

# curl 7.66+ waits for Retry-After between attempts.
curl --retry 5 --retry-max-time 600 \
  https://api.lahja.dev/v1/chat/completions \
  -H "Authorization: Bearer $LAHJA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "google/gemma-4-31b-it:arabic-gulf",
       "messages": [{"role": "user", "content": "هلا"}]}'

import os, time
from openai import OpenAI, APIStatusError

client = OpenAI(base_url="https://api.lahja.dev/v1",
                api_key=os.environ["LAHJA_API_KEY"], max_retries=0)

def chat(messages, model="google/gemma-4-31b-it:arabic-gulf", tries=6):
    for attempt in range(tries):
        try:
            return client.chat.completions.create(model=model, messages=messages)
        except APIStatusError as e:
            retryable = e.status_code == 503 or (
                e.status_code == 429 and e.type == "rate_limit_error")
            if not retryable or attempt == tries - 1:
                raise
            time.sleep(int(e.response.headers.get("retry-after", "10")))

const sleep = (s) => new Promise((r) => setTimeout(r, s * 1000));

async function chat(body, tries = 6) {
  for (let i = 0; ; i++) {
    const r = await fetch("https://api.lahja.dev/v1/chat/completions", {
      method: "POST",
      headers: {
        "Authorization": `Bearer ${process.env.LAHJA_API_KEY}`,
        "Content-Type": "application/json",
      },
      body: JSON.stringify(body),
    });
    if (r.ok) return r.json();
    const { error } = await r.json().catch(() => ({}));
    const retryable = r.status === 503 ||
      (r.status === 429 && error?.type === "rate_limit_error");
    if (!retryable || i === tries - 1) throw new Error(error?.message ?? r.statusText);
    await sleep(Number(r.headers.get("retry-after") ?? 10));
  }
}