Our scrubbing & audit layer, running Sigmix-authored firmware — with an optional PCIe card for the appliance.
Explore Sigmix Guard
Every request served on-box, sealed from the internet. KSA Cloud runs elastic and in-Kingdom.
See your postureSelf-serve keys open the two Gen-AI tools to your code: KSA Sovereign Gen-AI, processed in-Kingdom on Cohere Command A (Oracle, Riyadh), and Gen-AI Global, whose frontier models run outside the Kingdom. Every call draws on your workspace credits and obeys the same plan limits as the app.
// self-serve keys from the dashboard · one per tool, shown once
# KSA Sovereign Gen-AI — processed in-Kingdom curl https://ksa.sigmix.ai/api/chat \ -H "x-sigmix-key: $SIGMIX_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "message": "لخّص هذا التقرير" }'
Each tool is its own service with its own base URL, and each key belongs to one tool. The paths are the same on both — the base URL decides whether your call is sovereign.
https://ksa.sigmix.ai/api
Processed in-Kingdom on Cohere Command A (Oracle, Riyadh).
https://genai.sigmix.ai/api
Frontier models — Claude on AWS Bedrock in the US, and Google Gemini. Labelled Global everywhere: your request leaves the Kingdom.
Create keys yourself on the API keys page of your dashboard — no ticket, no sales call. Each key belongs to one person and one tool, and can be rotated or revoked without touching the others. Send it in the x-sigmix-key header on every request; entitlement, plan limits and credits are still checked on every call.
import os, requests resp = requests.post( "https://ksa.sigmix.ai/api/chat", headers={"x-sigmix-key": os.environ["SIGMIX_API_KEY"]}, json={"message": "Summarise Q3 revenue"}, timeout=120, ) data = resp.json() print(data["text"]) # the answer print(data["creditsCharged"]) # what this call cost
The same three paths on both tools. Each request returns one JSON response with the whole answer — not a stream.
| Method | Path | What it does |
|---|---|---|
| POST | /api/chat | A chat completion, in Arabic or English. Send message, or a messages array of user and assistant turns; model and conversationId are optional. Returns the answer, the tokens used and the credits charged. |
| GET | /api/models | The models this tool serves — on KSA Sovereign Gen-AI, in-Kingdom models only. |
| POST | /api/credits | Your remaining credit balance. |
// full reference, schemas and error codes live in the developer docs — link from your dashboard once you have a key
The base URL is the choice. KSA Sovereign Gen-AI is processed in-Kingdom. Gen-AI Global runs outside the Kingdom and is labelled so.
GET /api/models returns what this tool serves. On KSA Sovereign Gen-AI, that is in-Kingdom models only.
Entitlement, plan caps and credits are enforced on every API call exactly as in the workspace. The API is not a way around them.
Each call reserves its maximum cost first and settles only what it used. A call that fails, or is refused, is not charged.
API traffic is never used to train any model, ours or a vendor's.
There is no SDK to install: the API is plain JSON over HTTPS. Use whatever your team already runs, and keep the key on the server — never in a browser or an app bundle.
Limits come from your plan — requests and tokens per minute — and are the same in the app and over the API. Every refusal says why, and none of them is charged.
GET /api/models on the tool you are using. On KSA Sovereign Gen-AI it lists in-Kingdom models only. Leave model out and the tool's default is used.Create a key from your dashboard and make your first call. It draws on the same credits, and obeys the same limits, as your workspace.