VarunSH
OpenAI & Anthropic compatible

One API. Your AI Models.

Access your entire AI model infrastructure through one fast, reliable, developer-first API. Route, meter, and secure every request from a single endpoint.

5M free tokens / monthNo card required
POST /v1/chat/completions
curl https://api-gw.aimasteryedu.in/v1/chat/completions \
  -H "Authorization: Bearer $VARUNSH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "varunsh/atlas-pro",
    "messages": [
      { "role": "user", "content": "Explain routing in one line." }
    ]
  }'
200 OKmodel varunsh/atlas-prolatency 312mstokens 1,204req req_a1b2c3

Built to route across every major model ecosystem

OpenAIAnthropicGoogleDeepSeekQwenMistralxAIMetaMoonshotGLMMiniMaxNVIDIAOpenAIAnthropicGoogleDeepSeekQwenMistralxAIMetaMoonshotGLMMiniMaxNVIDIA

Why VarunSH

Infrastructure that gets out of your way

Everything you need to run models in production, without stitching together a dozen provider SDKs.

Intelligent routing

Priority, health, and fallback routing across provider endpoints — configured per model, not hardcoded.

Secure by default

Provider credentials stay server-side, encrypted at rest. SSRF-guarded custom endpoints. Scoped API keys.

Quota & rate limits

Per-plan token allowances and Redis-backed rate limiting, enforced server-side on every request.

Usage you can trust

Metered on real provider token counts, with a documented estimator fallback. Nothing fabricated.

Provider-agnostic

Add models and providers from the admin panel. The frontend and gateway never need a rewrite.

Streaming-first

Server-sent events with correct usage capture on stream completion, client disconnect handling, and cancellation.

Developer workflow

From signup to first request in minutes

01

Sign up

Create an account on the free plan. No card required.

02

Create an API key

Generate a scoped key from your dashboard. Shown once, stored hashed.

03

Point your SDK

Swap the base URL to VarunSH. Keep your existing OpenAI or Anthropic client.

04

Ship

Make your first request. Watch usage, latency, and routing in real time.

Simple pricing

Start free. Scale when you're ready.

Begin with 5M tokens a month at no cost. Move to daily allowances from 30M to 500M tokens as your workloads grow — all quotas configurable, all enforced server-side.

Free

5M / mo

Starter

30M / day

Growth

100M / day

Scale

500M / day

FAQ

Questions, answered

A unified API gateway that routes requests to your AI models through one OpenAI- and Anthropic-compatible endpoint. Point your existing SDK at our base URL and go.

Rarely. If you already use the OpenAI or Anthropic SDK, you only change the base URL and API key. The request and response shapes are compatible.

Usage is metered on actual tokens returned by the upstream model. Each plan includes a token allowance; pricing per model is transparent on the model's detail page.

No. By default we record only metadata needed for usage accounting — tokens, latency, status, model. Prompt and response bodies are never persisted.

When fallback routing is enabled for a model, requests automatically retry against the next healthy provider route in priority order, then return a normalized response.

Yes. The platform is provider-agnostic. An administrator adds a provider, configures its endpoint and credentials, maps a model, and it's live — no frontend changes.

Ship on one API today

Create your account, generate a key, and make your first request before your coffee gets cold.