How to Get an AI API Key in India

Getting an AI API key that you can actually put on a company card in India has historically meant an international card, a foreign biller, and an invoice your finance team can't easily reconcile. unoblox skips all of that: sign up, verify, generate a key, and you're calling GPT, Claude, DeepSeek, Qwen, Llama and more within minutes — billed in rupees on one monthly GST invoice.

Step by step: getting your key

1. Go to https://unoblox.ai/sign-in and create an account.
2. Verify your email address.
3. Open the API keys section of your dashboard and generate a new key. It will look like ub-gw- followed by a long random string.
4. Copy the key immediately — for security, the full value is only shown once at creation time.
5. Add balance to your account in rupees (no international card required), or use the free-tier model to start testing before you add any balance at all.

What the key looks like and how you use it

Every key follows the same ub-gw-... format regardless of which model you plan to call. You use it exactly like you would an OpenAI API key: as a bearer credential in the Authorization header, against the unoblox base URL.

base_url: https://api.unoblox.ai/v1
api_key: ub-gw-xxxxxxxxxxxxxxxxxxxx
model: qwen/qwen3-1.7b

Because the endpoint is OpenAI-compatible, any existing OpenAI SDK, LangChain integration, or raw HTTP client that already knows how to call an OpenAI-shaped API will work here with just the base URL and key swapped.

Keeping your key safe

• Store it as an environment variable in your deployment (for example, a variable named UNOBLOX_API_KEY) rather than hardcoding it into source files.
• Never commit a key to a public or shared repository.
• Generate a separate key per project or environment (development versus production) so you can revoke one without affecting the others.
• If a key is ever exposed, revoke it from the dashboard and generate a new one immediately — old keys stop working the moment they're revoked.

Billing once you start calling models

Every model you call — whether it's the free Qwen3 1.7B or a paid model like GPT-4o or Claude Opus — is metered per token and rolled into one monthly GST invoice from an Indian entity. There's no per-provider bill, no international card, and the invoice is claimable as input tax credit like any other GST-billed business expense.

Picking your first model to call

If you're just testing your integration, start with qwen/qwen3-1.7b — it's free (₹0) so you can validate your request format, error handling, and streaming setup without spending anything. Once that works, swap the model field to whichever paid model fits your task and check its live rate on /models.

Frequently asked questions

Do I need an international credit card to get an API key on unoblox? No. Sign-up, verification, and adding balance are all designed around rupee payment methods — no international card is required at any step.

How long does it take to get a working key? Sign-up and email verification typically take a few minutes, and the key is generated instantly once you're in the dashboard — you can be making your first call the same session.

Can I have more than one API key? Yes — most teams generate separate keys per project or environment so keys can be rotated or revoked independently.

What happens if I lose my key after creation? The full key value is shown only once, at creation time, for security. If you didn't save it, revoke it from the dashboard and generate a new one — you can't retrieve the original value later.

Is there a cost just to hold an API key? No — you're only billed for the tokens you actually use when calling a model. Holding an unused key costs nothing.

Which models can I call with one key? Every model on the unoblox catalog — GPT, Claude, DeepSeek, Qwen, Llama, Gemma, Kimi, Mistral, and Nemotron variants — is reachable with the same key against the same base URL.

Get started in rupees → https://unoblox.ai/sign-in