n8n AI automation with ₹-native pricing

Use n8n AI nodes with an India gateway

n8n's OpenAI-compatible AI node (under Tools → AI) accepts any custom endpoint and key. Point it at unoblox and run text generation, summarization, and classification at ₹-native rates, invoiced monthly from an Indian entity.

2-minute setup

1. Get an API key: unoblox.ai/sign-in → copy ub-gw-…
2. In n8n, add an OpenAI-compatible AI node (or an HTTP Request node for full control):
   • Base URL: https://api.unoblox.ai/v1
   • API Key: paste ub-gw-…
   • Model: qwen/qwen3-235b-a22b-instruct-2507 (or deepseek-ai/deepseek-v3.2)
3. Test with any text input; check the request and its ₹ cost on your dashboard

Raw HTTP Request node (if you need full control)

POST https://api.unoblox.ai/v1/chat/completions
Headers: Authorization: Bearer ub-gw-YOUR_KEY, Content-Type: application/json
Body:
{
  "model": "deepseek-ai/deepseek-v3.2",
  "messages": [{"role": "user", "content": "<map this to your input field via n8n's expression editor>"}]
}

Use this when you need fields — response format, temperature, tools — that the built-in AI node doesn't expose.

Which model for which n8n step

Use case	Best model	₹ per 1M (in / out)
Workflow summarization	Qwen3 235B A22B	₹9.07 / ₹55.44
Customer data classification	DeepSeek V3.2	₹26.21 / ₹38.30
Content generation	Qwen3 Max	₹120.95 / ₹604.77
Prototyping, low-stakes steps	Qwen3 1.7B	FREE

Real automation examples

• CRM enrichment: extract company names from inbound emails (Qwen), write to Postgres, one line item on the monthly invoice
• Support triage: tag tickets by urgency (DeepSeek), auto-assign to the right Slack channel
• Content ingestion: summarize PDFs on the free Qwen3 1.7B tier, store vectors in Pinecone or Qdrant for later search
• Lead scoring: classify inbound form submissions, push qualified leads to a CRM node, park the rest in a review sheet

FAQ

Do I keep n8n's own licensing? Yes. n8n self-hosted or n8n Cloud stays exactly as it is — you're only swapping the AI node's backend endpoint and key.

How often does unoblox refresh its model catalog? New models are added as they ship upstream — GPT, Qwen, DeepSeek updates and more. Check the live list at /models.

Can I combine n8n conditions with unoblox calls in one workflow? Yes. unoblox is just the AI node; every n8n branch, HTTP, database, and conditional node works around it unchanged.

Will this change how I get invoiced? Yes, for the simpler: instead of several AI subscriptions, you get one ₹-denominated GST invoice covering every model you used that month.

How do I know if a model supports streaming? All unoblox chat models support streaming. Check a model's feature flags at /models/{slug}.

Can I set a system prompt from within n8n? Yes. The AI node's prompt field maps to the system role, and the HTTP Request node lets you set it explicitly in the JSON body.

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