Make.com AI with unoblox (India)

Make.com's dedicated OpenAI app only ever talks to OpenAI's own servers — there's no field to repoint it. The generic HTTP module, however, can call any REST endpoint, and unoblox implements the same chat-completions contract OpenAI does, just billed in rupees on a monthly GST invoice. That's the module Indian teams use to wire GPT, Claude, DeepSeek, Qwen or Llama into a no-code scenario without a personal international card behind it.

Build the HTTP module call

Add an HTTP > Make a request module with:

• Method: POST
• URL: https://api.unoblox.ai/v1/chat/completions
• Headers: Authorization: Bearer ub-gw-xxxxxxxxxxxxxxxxxxxxxxxxxxxx and Content-Type: application/json
• Body type: raw JSON, for example:

{
  "model": "deepseek-ai/deepseek-v4-flash",
  "messages": [
    {"role": "user", "content": "Summarise the incoming record in one sentence."}
  ]
}

Replace the static content text with a field mapped from an earlier module (a webhook payload, a spreadsheet row, a form submission) the same way you'd map any other HTTP body value in Make.

Parsing the response

Toggle Parse response on the HTTP module so the reply's choices[].message.content becomes a mappable output field for the rest of your scenario — write it to a sheet, post it to Slack, or send it as an email body.

Choosing a model for automation scenarios

Model id	₹ / 1M input	₹ / 1M output
deepseek-ai/deepseek-v4-flash	₹9.07	₹18.14
qwen/qwen3-235b-a22b-instruct-2507	₹9.07	₹55.44
google/gemma-4-31b-it	₹13.10	₹38.30
qwen/qwen3-1.7b (free)	₹0.00	₹0.00

For anything else in the catalog — GPT-5 nano, Claude, Mistral Small, Nemotron and more — check the live ₹ rate on /models rather than assuming it matches a row here.

Cost and governance for no-code teams

Make scenarios often fire per row or per incoming webhook, which can run into the thousands over a month. Because unoblox meters by token rather than by seat or by request, a scenario's AI spend is visible and auditable against actual usage, and it settles under one company invoice instead of being spread across individual team members' personal accounts.

Frequently asked questions

Can I use Make's built-in OpenAI app instead? No — that app only targets OpenAI's own host; the generic HTTP module is what lets a scenario call unoblox.

Does this work inside a loop over many records? Yes — place the HTTP module inside an iterator or repeater and it fires once per array item, same as any other HTTP call in Make.

Where should the key live in a shared scenario? Store it as a Make connection or a protected scenario variable rather than pasting it directly into a module that might later be exported or duplicated.

Which model is cheapest for short summarisation scenarios? DeepSeek V4 Flash and Qwen3 235B-A22B are the least expensive options in the table above for that kind of task.

Can I stream the response into a scenario? No — Make's HTTP module reads a full response rather than a token stream, so leave stream unset or false and read the completed reply.

How is this billed? A monthly GST invoice in rupees from an Indian entity, input tax credit claimable, with no international card involved.

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