Migrate from OpenRouter (India)

Why OpenRouter users in India switch to unoblox

OpenRouter's model marketplace is convenient — one API, many providers — but it settles internationally, so Indian teams absorb payment friction and can't claim GST input-tax credit on the spend. unoblox offers a comparable range of models (OpenAI, Anthropic, DeepSeek, Qwen, Llama, Gemma) through one ₹-billed endpoint, invoiced monthly by an Indian entity.

The migration path

Step 1 — Sign up at unoblox.ai and copy your ub-gw-… key from the admin console.

Step 2 — Update your code:

# OLD (OpenRouter)
from openai import OpenAI
client = OpenAI(
    api_key="sk-or-v1-...",
    base_url="https://openrouter.io/api/v1"
)

# NEW (unoblox)
client = OpenAI(
    api_key="ub-gw-...",
    base_url="https://api.unoblox.ai/v1"
)

Step 3 — Test, then flip traffic. Run a handful of requests against the new endpoint, confirm output and latency look right, then move everything over. Because the request and response schema are unchanged, most teams complete this in an afternoon rather than a sprint.

₹ pricing, no surprises

Model	₹ input / 1M	₹ output / 1M
GPT-4o	₹252	₹1008
Claude Opus	₹504	₹2520
DeepSeek V3.2	₹26.21	₹38.30
Qwen3 Max	₹120.95	₹604.77
Qwen3 1.7B	₹0	₹0

Rates are published per model and don't move with a foreign-exchange market — what you see on /models is what lands on your invoice.

Why unoblox wins on the finance side

• GST invoice, every month. Claim your input-tax credit on the full spend, on one document.
• ₹-only, always. No international payment method, no reconciling a foreign-currency line item against a rupee budget.
• Live model listing. /models shows real-time availability and current ₹ rates — no stale price sheet.
• Same model breadth. OpenAI, Anthropic, and open-source models are all present, alongside India-relevant options like Qwen3.
• Predictable settlement. Invoiced monthly, with payment terms negotiable at volume — no prepaid credit balance to babysit.
• Same request parameters. temperature, top_p, stream, and tool-calling schemas all pass through unchanged — nothing to remap in your application code.

Frequently asked questions

Can I test both providers side by side? Yes — add unoblox to your router config, send a percentage of traffic there, and watch error rates and latency before committing fully.

How does billing differ from OpenRouter's credit system? unoblox invoices monthly instead of running on prepaid credits, so there's no balance to top up before a weekend traffic spike.

Can I export my OpenRouter usage history? Not in bulk from OpenRouter itself — pull anything you need before cutting over, since new usage logs start fresh on unoblox.

Do you support high-volume batch requests? Yes — loop your requests and they queue fairly. It's a common pattern for overnight data generation or dataset labeling.

How do I check live model availability? Call /models from your app; it returns current status and ₹ pricing so you can handle a temporarily unavailable model gracefully.

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