Zapier + AI API: automation in rupees

Run Zapier's AI actions with ₹-native pricing

Zapier doesn't need a special unoblox integration — "Webhooks by Zapier" (POST JSON) can call any OpenAI-compatible endpoint directly. Point it at unoblox and get an India-billed gateway: one GST invoice in rupees, no international card.

Quick 3-step connection

1. Sign up at unoblox.ai/sign-in, copy your API key ub-gw-…
2. In Zapier, add a Webhooks by Zapier → POST action:
   • URL: https://api.unoblox.ai/v1/chat/completions
   • Payload type: JSON
   • Headers: Authorization: Bearer ub-gw-… and Content-Type: application/json
   • Data: {"model": "qwen/qwen3-235b-a22b-instruct-2507", "messages": [{"role": "user", "content": "<text>"}]}
3. Test with sample data — the invoice line item appears on your dashboard automatically

Reading the response back into your Zap

Zapier's built-in Formatter step (or a second Webhooks "parse" step) can pull choices[0].message.content out of the JSON response as its own field, ready to feed into the next action — a Slack message, an Airtable row, a Gmail draft.

Popular automation workflows

Trigger	Action	Model	Typical spend
New Gmail	Summarize email	Qwen3 235B A22B	a few paise per email
Slack message	Classify sentiment	DeepSeek V3.2	a few paise per message
Google Form	Draft a response	Qwen3 Max	tens of paise per form
Airtable row	Extract and tag fields	Qwen3 1.7B	₹0 (free tier)

Costs above are illustrative and driven by prompt length — check /dashboard for your actual per-call spend.

Pricing: same ₹ rate for every use case

• Qwen3.8-27B ₹16.32 / ₹48.96 per 1M tokens — fast, handles vision
• DeepSeek V3.2 ₹26.21 / ₹38.30 per 1M tokens — deeper reasoning
• Qwen3 1.7B ₹0 per 1M tokens — free tier, rate-limited, monthly caps

See live rates at /models; billing is metered per token and invoiced monthly.

FAQ

Can I add a system prompt in Zapier? Yes. Extend the messages array in the Data field: {"role": "system", "content": "Act as X"} before the user message.

How long does Zapier wait for a response? Zapier's default webhook timeout is generous enough for most calls; unoblox responses typically arrive well within it. For long generations, add a Zapier Delay step and poll instead.

Can I parse the AI response inside Zapier? Yes. Use Zapier's Formatter (or a "Parse JSON" step) to extract choices[0].message.content as a variable other steps can use.

Do you charge per Zapier task or per token? Per token, always. Zapier's own task-based pricing is separate from unoblox's ₹-per-token billing for the model call itself.

Can I use unoblox with Zapier's native OpenAI integration instead of Webhooks? No — Zapier's built-in OpenAI action is hardcoded to api.openai.com. Webhooks by Zapier is the way to point at any OpenAI-compatible endpoint, including unoblox.

Is data encrypted in transit? Yes. Every call to api.unoblox.ai runs over TLS. Request bodies are not used for model training.

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