OpenAI-compatible API

AI models, including Arabic, served on community GPUs

Call chat models with the OpenAI SDK you already use — just change the base URL. Pay per token from your prepaid wallet, no subscription. Includes ALLaM, the Arabic model from SDAIA, which very few providers offer per token.

New accounts get free trial credit automatically the first time they open AI chat.

Try it in your browser Get an API key Earn with your graphics card

Models and prices

Model	Input / 1M tokens	Output / 1M tokens	Context	Status
Loading…

Prices in US dollars per 1 million tokens. You're billed for the exact tokens in each response's usage field.

Quick start

Create an account and add funds to your wallet (mada, Apple Pay or card).
Create an API key in Settings → Developers.
Point any OpenAI client at the base URL below and use a model id from the table.

cURL

curl https://cloudorbit.relon.uk/api/v1/chat/completions \
  -H "Authorization: Bearer $CLOUD_ORBIT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "humain-ai/allam-7b-instruct",
    "messages": [{"role": "user", "content": "Give three tips to save energy at home"}],
    "stream": true
  }'

Python (openai)

from openai import OpenAI

client = OpenAI(base_url="https://cloudorbit.relon.uk/api/v1", api_key="YOUR_CLOUD_ORBIT_API_KEY")
resp = client.chat.completions.create(
    model="humain-ai/allam-7b-instruct",
    messages=[{"role": "user", "content": "اشرح الحوسبة السحابية في سطرين"}],
)
print(resp.choices[0].message.content)
print(resp.usage)  # prompt / completion tokens you were billed for

Node.js (openai)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://cloudorbit.relon.uk/api/v1", apiKey: process.env.CLOUD_ORBIT_API_KEY });
const stream = await client.chat.completions.create({
  model: "humain-ai/allam-7b-instruct",
  messages: [{ role: "user", content: "Summarize the benefits of solar energy" }],
  stream: true,
});
for await (const chunk of stream) process.stdout.write(chunk.choices[0]?.delta?.content ?? "");

Data stays in Saudi Arabia — when you need it

Add "data_residency": "SA" to any request (or tick the box in AI chat) and it is processed only on devices connected from Saudi Arabia — never routed abroad. If none is free you get a clear 503 instead of a silent fallback. Costs 20% more; device owners share the premium.

How your data is handled

• Cloud Orbit does not store your prompts or the model's answers. They pass through our servers in memory only.
• We keep only billing metadata: model, token counts, time and which device served the request.
• Your data is never used to train models.
Requests run on computers owned by members of our community, whose owners could technically observe them. Don't send passwords, personal data or confidential material. A verified-devices tier for sensitive data is planned.

Limits

• Responses stream token by token; if a device doesn't pick up your request within a few seconds, it's moved to another device automatically.
• Each request can run for up to 3 minutes. Rate limits follow your API key settings.
• If no device is serving a model right now you'll get a 503 with code no_capacity — retry shortly.

AI models, including Arabic, served on community GPUs

New accounts get free trial credit automatically the first time they open AI chat.

Model	Input / 1M tokens	Output / 1M tokens	Context	Status
Loading…

Model

Input / 1M tokens

Output / 1M tokens

Context

Status

Loading…

Quick start

Create an account and add funds to your wallet (mada, Apple Pay or card).

Create an API key in Settings → Developers.

Point any OpenAI client at the base URL below and use a model id from the table.

cURL

curl https://cloudorbit.relon.uk/api/v1/chat/completions \
  -H "Authorization: Bearer $CLOUD_ORBIT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "humain-ai/allam-7b-instruct",
    "messages": [{"role": "user", "content": "Give three tips to save energy at home"}],
    "stream": true
  }'

Python (openai)

from openai import OpenAI

client = OpenAI(base_url="https://cloudorbit.relon.uk/api/v1", api_key="YOUR_CLOUD_ORBIT_API_KEY")
resp = client.chat.completions.create(
    model="humain-ai/allam-7b-instruct",
    messages=[{"role": "user", "content": "اشرح الحوسبة السحابية في سطرين"}],
)
print(resp.choices[0].message.content)
print(resp.usage)  # prompt / completion tokens you were billed for

Node.js (openai)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://cloudorbit.relon.uk/api/v1", apiKey: process.env.CLOUD_ORBIT_API_KEY });
const stream = await client.chat.completions.create({
  model: "humain-ai/allam-7b-instruct",
  messages: [{ role: "user", content: "Summarize the benefits of solar energy" }],
  stream: true,
});
for await (const chunk of stream) process.stdout.write(chunk.choices[0]?.delta?.content ?? "");

How your data is handled

• Cloud Orbit does not store your prompts or the model's answers. They pass through our servers in memory only.

• We keep only billing metadata: model, token counts, time and which device served the request.

• Your data is never used to train models.

Requests run on computers owned by members of our community, whose owners could technically observe them. Don't send passwords, personal data or confidential material. A verified-devices tier for sensitive data is planned.

Limits

• Responses stream token by token; if a device doesn't pick up your request within a few seconds, it's moved to another device automatically.

• Each request can run for up to 3 minutes. Rate limits follow your API key settings.

• If no device is serving a model right now you'll get a 503 with code no_capacity — retry shortly.