नभ /nʌbʱ/ · the skyNo Credit Card Needed

Every model under one sky.

Q: Do my prepaid credits expire?

Purchased credits never expire. Free-plan accounts also get 50 free credits automatically on the 1st of every month, on top of the signup bonus. Pro's 2,000 monthly credits roll over, up to 6,000, and end if you leave Pro.

India's prepaid LLM API marketplace. One API key, 21 models, billed in rupees. Llama, Mistral, Qwen, DeepSeek and 17 more, on OpenAI-compatible endpoints. Pay via UPI, RuPay, or net banking — no forex markup, no forced subscription, no international credit card required.

Start with 50 Free Credits Browse All 21 Models

No Credit Card Needed OpenAI-Compatible API Billed in Paise UPI & Indian Cards

Product SurfaceBuilt for Shipping

One OpenAI-compatible endpoint, prepaid credits, scoped keys and per-token billing — everything you need to go from key to production.

BillingPaise

KeysScoped

PaymentsUPI

SpecOpenAI

~/nabh

api.openai.com/v1

api.nabh.cloud/v1

One line. Same SDK, same streaming, same errors.

✓ cost 0.01 credits · balance 49.99 credits

~/nabh — curl

$ curl https://api.nabh.cloud/v1/chat/completions \

-H "Authorization: Bearer nbh_••••••••" \

-d '{ "model": "gpt-oss-20b", "stream": true,

"messages": [{ "role": "user",

"content": "Show the Python snippet to switch to this endpoint." }] }'

✓ done · cost 0.00 credits · balance 50.00 credits

Streaming BuildToken-by-Token

Tokens start moving before the route settles. No buffering, no faux typing.

FormatSSE Stream

BillingPer Token

Inventory21 Models

Chat, code, embeddings, vision, speech and reasoning behind one key.

LanguagesMultilingual

PaymentsUPI First

CreditsPrepaid Credits

Buy credits via UPI, cards, or net banking. Purchased credits never expire.

Minimum₹500

Free Credits50

PaymentsUPI

ExpiryNever

75 km

The catalog. Noctilucent clouds included.

21 models, one request shape. Most models share a single endpoint — swap a model string, nothing else moves. A few use their own model-scoped URL instead — same auth, same shape.

Real clouds stay in the troposphere. Around 75 km you are in the mesosphere, where noctilucent clouds can appear at night.Noctilucent clouds are real — they form around 75 km, in the mesosphere.

GPT OSS 20B

OpenAI

LIVE

Free tier

Reasoning

gpt-oss-20b delivers strong performance with fast inference and efficient reasoning, and is available on every plan, including Free.

₹0.01/1K in · ₹0.03/1K out

128K ctx

Qwen3.6 27B

Alibaba

LIVE

Chat · Vision · Function Calling

A 27B-parameter multimodal model from Alibaba's Qwen team, built as an image-text-to-text model with a 256K-token context window and function calling.

₹0.06/1K in · ₹0.39/1K out

256K ctx

Llama 3.3 70B Instruct

GPT OSS 120B

OpenAI

LIVE

pro plan

Reasoning

A cutting-edge model designed for high-level reasoning, instruction-following, and advanced agent capabilities — strong across both general-purpose and specialized applications.

₹0.02/1K in · ₹0.06/1K out

128K ctx

Qwen3.5 397B

Alibaba

LIVE

pro plan

Chat · Function Calling

Alibaba's largest Qwen3.5 model — 397B parameters and a 256K context window, built for the most demanding reasoning and multimodal tasks.

₹0.09/1K in · ₹0.52/1K out

256K ctx

Gemma 4 26B

Google

LIVE

Reasoning · Vision

Google DeepMind's Gemma 4 26B A4B — an instruction-tuned mixture-of-experts model with 25.2B total parameters but only 3.8B active per token, giving near-31B quality at MoE-efficient cost. Genuinely multimodal (text, images, and short video), with a 256K context window, native function calling, a toggleable reasoning mode, and structured output support. Scores 77.1% on LiveCodeBench v6 and 88.3% on AIME 2026. Apache 2.0. For vision requests, set max_tokens to at least 1000–1500 — the model spends part of its budget on internal reasoning before producing a visible answer.

₹0.02/1K in · ₹0.04/1K out

256K ctx

Kokoro TTS

Hexgrad

LIVE

Text to Speech

Kokoro is a lightweight, open-weight TTS model — just 82 million parameters on a decoder-only StyleTTS2 + ISTFTNet architecture, no diffusion. Despite its size it delivers quality comparable to much larger TTS models while running significantly faster and cheaper, with 54 voices across 8 languages. Trained exclusively on permissive, non-copyrighted audio. Apache 2.0.

₹0.15/1K chars

Kimi K2.7 Code

Moonshot AI

LIVE

Code · Vision · Function Calling

Moonshot AI's Kimi K2.7, tuned for code.

₹0.13/1K in · ₹0.52/1K out

256K ctx

See All 21 Models & Live Latency →

The PlatformBilling, Streaming & Access →

Prepaid credits billed in paise, real SSE streaming, one-line migration, scoped keys, an invoice for every recharge.

QuickstartFive Minutes to First Token →

Swap two lines in the OpenAI SDK — base URL and key. Same shapes, same streaming, same errors.

45 km

Billing, streaming, and API keys — built into the platform.

Everything between your code and the GPU is ours to obsess over — billing in paise, streaming that actually streams, keys you can rotate without downtime.

Your Prepaid Balance

Prepaid — like your phone. Buy credits over UPI, watch every request deduct its exact cost, and credits you buy never expire.

Balance1,240.50credits

UPI Linked

+₹500+₹2,000+₹5,000

See Full Pricing & Recharge →

Real-Time Token Streaming

Tokens arrive as the model emits them — no proxy buffering in between.

Real SSE streaming, chunk-by-chunk

One Line to Migrate

OpenAI-compatible surface. Point your existing SDK at us and ship.

base_url="https://api.openai.com/v1"

base_url="https://api.nabh.cloud/v1"

Built for India

Billed in rupees on the payment rails Indian teams already use.

UPIRuPayNet BankingInvoices

Grown-Up API Keys

Scoped keys with IP allowlists and zero-downtime rotation.

ip_allowlistauto-expirygrace rotation

Read the Docs →

30 km

Compliance and billing, built in.

Billing, access, retention and audit trails sit beside the model catalog, not behind it.

Privacy

Memory, on Your Terms

Prompts and completions are processed to deliver your request, bill usage, and prevent abuse — never for training. API traffic is never stored as memory; persistent memory is an app-only feature you can view, delete, or turn off anytime.

Billing

Invoices for Every Recharge

Credit purchases stay clean for Indian finance teams and procurement flows.

Access

Scoped Keys

IP allowlists, expiry windows and rotation flows are first-class console controls.

Compliance

DPDP-Ready Posture

Export your data or delete your account from the console — both run as first-class flows, not support tickets.

Key Rotation24h Grace

Invoice CurrencyINR

Data ExportCSV/JSON

Admin Access2FA Gated

Every operational promise on this page maps to a screen in the console.

Read the Security & Access Docs →

15 km

Five minutes to first token.

If you've used the OpenAI SDK, you already know our API. Same shapes, same streaming, same errors — different bill.

Sign Up, Get 50 Free Credits

Email or Google. Credits land before your coffee brews.

Mint a Key

nbh_… — scope it, allowlist it, or don't. Your call.

Swap the Base URL

Responses carry X-Balance-INR so finance stops asking.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.nabh.cloud/v1",
    api_key="nbh_...",  # that's the whole migration
)

stream = client.chat.completions.create(
    model="gpt-oss-20b",
    messages=[{"role": "user", "content": "नमस्ते!"}],
    stream=True,
)

for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.nabh.cloud/v1",
  apiKey: process.env.NABH_API_KEY,
});

const stream = await client.chat.completions.create({
  model: "gpt-oss-20b",
  messages: [{ role: "user", content: "नमस्ते!" }],
  stream: true,
});

for await (const chunk of stream)
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");

curl https://api.nabh.cloud/v1/chat/completions \
  -H "Authorization: Bearer $NABH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-oss-20b",
    "stream": true,
    "messages": [{ "role": "user", "content": "नमस्ते!" }]
  }'

# data: {"delta":{"content":"नमस्ते"}}   ← arrives immediately
# data: {"delta":{"content":"!"}}      ← and keeps arriving

Read the Full Documentation →

2 km

Prepaid. Like your phone.

Tokens cost the same on every plan — plans only raise your ceilings. Load up on nabh credits, spend them down, top up when you like.

How It Works: 1 nabh credit = ₹1. Every request deducts credits at the per-model rate below. Plans change rate limits and modality access, never the token price.

Free

₹0/mo

Start Exploring

50 Free Credits on Signup, Then 50 Every Month
Access: GPT OSS 20B (Free Tier)
5 Requests/Min Rate Limit
500 Requests/Day
1 API Key
7-Day Analytics History
Self-Serve Docs & FAQ

Get 50 Free Credits

Developer

Pay-as-You-Go

No Monthly Fee

Access to 17 Models
60 Requests/Min Rate Limit
10,000 Requests/Day
5 API Keys
30-Day Analytics History
Email Support
Team Seat Management
Recharge via UPI/RuPay/Cards

Start Building

Pro

~~₹2,000~~₹999/mo

For Teams & Scale

Includes 2,000 Credits Every Month
Unused Credits Roll Over, Up to 6,000
All 21 Models
120 Requests/Min Rate Limit
Unlimited Daily Requests
50 API Keys
12-Month Analytics History
Priority Email Support
Team Seat Management
Custom Recharge Limits

Go Pro

Enterprise

Custom

For Regulated Teams

Dedicated Capacity
Custom Rate Limits
100 API Keys
DPDP Agreements & Custom Invoicing
Direct Line to Our Team

Talk to Us

See Full Pricing, Sample Rates & Cost Calculator →

0 km

Questions

Do my prepaid credits expire? +

Purchased credits — never. Free-plan accounts also get 50 free credits automatically on the 1st of every month, on top of the signup bonus; whatever you bought yourself stays until you spend it. Pro's 2,000 monthly credits roll over too, up to 6,000, and end if you leave Pro.

Is this really OpenAI-compatible? +

Request and response shapes, streaming format, and error codes follow the OpenAI spec. Existing SDKs work by changing base_url and the key. If something behaves differently, that's a bug — tell us.

Do I pay anything on top of the credits? +

No. Every ₹1 you pay buys 1 credit, and every recharge comes with an invoice. Per-request rates shown are exactly what's deducted from your credits — no surprise line items.

Is my data used for training? +

Your prompts and completions are processed only to serve your request, bill usage, and prevent abuse — never for training. Usage logs never store the text itself. API traffic has no persistent memory: nothing from an API request is extracted or kept as a durable fact. Persistent memory is a nabh.cloud app feature only, and you can view, delete, or turn it off in Settings. See our Privacy Policy for how data is handled.

Which payment methods work? +

Any UPI app (GPay, PhonePe, Paytm), RuPay, Visa, Mastercard, and net banking via Razorpay. Enterprise can pay by invoice.

Can I use nabh.cloud without an international credit card? +

We bill in INR and accept UPI, RuPay, Indian net banking, and domestic cards. No forex, no international credit card needed.

Start building on नभ.

50 free credits are waiting. Billing is per request, in paise. See you up here.

Get Your API Key Read the Docs