Skip to content
AI API Gateway · India

The AI API gateway for India.

One OpenAI-compatible endpoint for every major model — GPT, Claude, DeepSeek, Qwen, Llama, Kimi. Billed in rupees, on a GST invoice, with per-key budgets and full usage visibility.

OpenAI-compatible · ₹-native billing · GST invoice · No international card

Why a gateway

One control surface for all your AI spend

Stop juggling a dozen provider consoles, international cards and scattered keys. Route everything through one Indian gateway and get cost, security and governance in the box.

Cost control

Give every API key a hard rupee budget and draw from one shared credit pool. A leaked key can't drain the org balance.

Usage visibility

Every request is metered in rupees — by key, by model, by workspace. See exactly where spend goes, in ₹, not guesswork.

Key security

Keys are SHA-256 hashed at rest and scoped per team. Upstream provider keys are never exposed to your users.

Model governance

Allow or block models and providers per key. Standardise what each team can call — and bill it all on one invoice.

Drop-in replacement

Change two lines. Keep your code.

unoblox speaks the OpenAI API your stack already uses. Point the base URL at the gateway, swap in a ub-gw- key, and pick any model in the catalog — streaming, tool calls and JSON mode all work unchanged.

  • Same SDKs — OpenAI, LangChain, LlamaIndex, Vercel AI SDK — no rewrite.
  • Same request shape — /v1/chat/completions, /v1/responses and an Anthropic-style /v1/messages.
  • Switch models by id — openai/gpt-5.2 → deepseek-ai/deepseek-v4-flash in one field.
curl
curl https://api.unoblox.ai/v1/chat/completions \
  -H "Authorization: Bearer $UNOBLOX_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5.2",
    "messages": [{ "role": "user", "content": "Draft a GST invoice email." }]
  }'
One catalog

Every major model, one bill

Reach frontier and open models across seven provider families through a single key — with live rupee pricing on every row.

OpenAI

GPT-5.x, GPT-4.1, o-series

Anthropic

Claude Opus, Sonnet, Haiku

DeepSeek

V4 Flash, V4.1 Flash, V3.2

Alibaba Qwen

Qwen3 Max, 235B, 3.8-27B, 1.7B (free)

Meta Llama

Llama 4 Maverick, Scout

Moonshot

Kimi K2.7 Code

…and more

One catalog, live ₹ pricing

₹-native billing

Billed in rupees, properly

  • Prepaid ₹ wallet — One rupee balance for the whole org; top up or auto-top-up.
  • Monthly GST invoice — From an Indian entity, so procurement and finance are happy.
  • Input tax credit — Claim ITC on your AI spend where your business is GST-registered.
  • Purchase orders — POs and standard terms — no international card needed.
How it works

Live in an afternoon

  • 1

    Create a key

    Sign in, fund a rupee wallet, and mint a ub-gw- key with a budget and a model allow-list.

  • 2

    Send requests

    Point your OpenAI SDK at api.unoblox.ai/v1. Streaming, tools and JSON mode work unchanged.

  • 3

    Track & control

    Watch ₹ usage per key and model, set budgets and alerts, rotate keys — all from one console.

Beyond chat, the same key reaches native jev decisions.

FAQ

AI API gateway, answered

What is the unoblox AI API gateway?

A single OpenAI-compatible API in front of every major model. You call one endpoint with one key; unoblox routes to the provider, meters usage in rupees, and gives you one GST invoice.

Is it really OpenAI-compatible?

Yes. Point your existing OpenAI SDK at https://api.unoblox.ai/v1 with a ub-gw- key. Chat completions, streaming, tool calls and JSON mode work unchanged — switch model by changing the model id.

Which models can I call?

Many models across OpenAI (GPT-4.1/5.x), Anthropic (Claude), DeepSeek (V3.2/V4), Alibaba Qwen (3.x), Meta Llama and Moonshot Kimi. The live list and ₹ prices are on the models page.

How am I billed?

In rupees, from a prepaid ₹ wallet, on a monthly GST invoice from an Indian entity. GST-registered businesses can claim input tax credit in the normal way. No international card is required.

Can I cap spend per team or per key?

Yes. Every key can carry a rupee budget and an allow-list of models; usage is recorded in rupees per key and per model, drawn from one shared credit pool.

Where does inference run?

On the model provider's infrastructure, and for select open models on unoblox-operated GPUs in India. unoblox provides the gateway, the Indian contract, ₹ billing and the controls.

Start building in rupees

Create a key, call every model through one endpoint, and pay only for what you use — in ₹, on a GST invoice.