USA Tech Solutions AI API
OpenAI-compatible · Prepaid · INR and USD

One API key for frontier models. Prepaid, pay in INR or USD.

Drop-in replacement for the OpenAI API. Use GPT, Grok, Kimi and (soon) Claude through a single base_url with your existing openai SDK. Top up credits in INR via Razorpay or in USD via PayPal, and pay only for the tokens you use. Built for developers in India, South Asia and Africa who want frontier models without a foreign card.

Streaming Tool calling Embeddings Per-key rate limits No negative balance

Models and prices

Prices are USD per 1M tokens. Live list: GET /v1/models. Reasoning tokens are billed as output tokens.

Model idProviderContextInput $/1MOutput $/1MStatus
gpt-chat-latestOpenAI128K$6.00$25.00Available
grok-4.7xAI256K$7.50$37.50Available reasoning tokens billed as output
kimi-k3.1-proMoonshot256K$3.00$12.00Available reasoning tokens billed as output
text-embedding-3-smallOpenAI8K$0.05—Available POST /v1/embeddings
claude-opus-5-5Anthropic200K$37.50$187.50Coming soon awaiting Anthropic access on Azure
claude-fable-5-1Anthropic200K$10.00$50.00Coming soon
claude-fable-5Anthropic200K$10.00$50.00Coming soon
claude-sonnet-5Anthropic200K$7.50$37.50Coming soon
deepseek-v3.2DeepSeek128K$1.50$6.00Coming soon

Credit packs

Prepaid. Pay in INR via Razorpay (live USD→INR rate) or USD via PayPal. Credits are deducted per request from the token usage the model reports.

$5
Trial

Try every model with a small top-up.

$20
Starter

For side projects and prototypes.

$50
Pro

For apps in production.

$100
Scale

For teams and higher volume.

Sign up and add credits

Quickstart

Base URL: https://api.usatechsolution.ai/v1 (custom domain api.usatechsolution.ai coming). Create a key in the dashboard, then use it as a Bearer token. Any OpenAI SDK works by changing base_url.

curl https://api.usatechsolution.ai/v1/chat/completions \
  -H "Authorization: Bearer usk_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-chat-latest",
    "messages": [{"role": "user", "content": "Say hello in Hindi"}],
    "max_tokens": 200
  }'

# Streaming: add "stream": true and read the text/event-stream response.
# List models:
curl https://api.usatechsolution.ai/v1/models -H "Authorization: Bearer usk_live_..."
pip install openai

from openai import OpenAI

client = OpenAI(
    base_url="https://api.usatechsolution.ai/v1",
    api_key="usk_live_...",
)

# Non-streaming
resp = client.chat.completions.create(
    model="gpt-chat-latest",
    messages=[{"role": "user", "content": "Say hello in Hindi"}],
    max_tokens=200,
)
print(resp.choices[0].message.content)
print(resp.usage)  # prompt_tokens / completion_tokens billed

# Streaming
stream = client.chat.completions.create(
    model="grok-4.7",
    messages=[{"role": "user", "content": "Write a haiku about monsoon"}],
    stream=True,
)
for chunk in stream:
    delta = chunk.choices[0].delta.content if chunk.choices else None
    if delta:
        print(delta, end="", flush=True)

# Embeddings
emb = client.embeddings.create(model="text-embedding-3-small", input="hello world")
print(len(emb.data[0].embedding))
npm install openai

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.usatechsolution.ai/v1",
  apiKey: "usk_live_...",
});

const completion = await client.chat.completions.create({
  model: "kimi-k3.1-pro",
  messages: [{ role: "user", content: "Explain prepaid API credits in one line" }],
  max_tokens: 300,
});
console.log(completion.choices[0].message.content);

// Streaming
const stream = await client.chat.completions.create({
  model: "gpt-chat-latest",
  messages: [{ role: "user", content: "Count to five" }],
  stream: true,
});
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
}

Errors you may see

HTTPcodeMeaning
401invalid_api_keyMissing, wrong or revoked key.
402insufficient_creditsBalance is empty. Add credits in the dashboard.
429rate_limit_exceededOver the per-key limit (requests/min or daily cap). Retry after the retry-after header.
400model_not_foundUnknown or not-yet-available model id. See GET /v1/models.
403key_suspendedKey suspended for abuse or on request. Contact support.

FAQ

What is this?

A pay-as-you-go gateway that resells frontier models hosted on Azure AI Foundry through one OpenAI-compatible endpoint. You get one API key, one balance, and one base_url for every model listed above.

How do credits work?

Credits are prepaid. For every request we read the token usage reported by the model and deduct prompt tokens × input price + completion tokens × output price from your balance. Deductions are atomic, so your balance can never go negative. When it reaches zero, requests return HTTP 402 until you top up.

What are the rate limits?

Per key: 60 requests per minute, a maximum of 8,192 output tokens per request (larger max_tokens values are clamped), and a default spend cap of $20 per key per day. Contact us to raise these for production workloads.

How do I pay?

In INR via Razorpay (UPI, cards, net banking) using a live USD→INR rate, or in USD via PayPal. Razorpay credits land automatically; PayPal payments are credited manually within 24 hours after you email your receipt to anylohar49@gmail.com.

Can I get a refund?

Unused credits are refundable within 7 days of purchase on request. Email us with your account email and payment reference.

Is my data used for training?

No. Requests pass through to Azure AI Foundry and are not used to train models. We log request metadata (timestamps, model, token counts, IP) for billing and abuse prevention; we do not store prompt or response content.

When will Claude be available?

The Claude deployments are provisioned and the adapter is built; we are waiting on Anthropic's access approval for Claude on Azure. They will appear in GET /v1/models as soon as access is granted.

Support

Email anylohar49@gmail.com. Include your account email and, for API issues, the model id and approximate time of the request.