Change one base URL in your OpenAI or Anthropic SDK and get every frontier and open-weight model through one prepaid key: transparent per-host prices, measured uptime and latency, automatic failover, hard budgets, and invoices your accountant can file.
Prepaid credits, no subscription needed. INR and USD.
from openai import OpenAI
client = OpenAI(base_url="https://api.ai.ml/v1", api_key="aiml-live-…")
client.chat.completions.create(model="openai/gpt-5-mini", messages=[{"role": "user", "content": "hi"}])Top up in INR by UPI, net banking or card at the day's RBI rate plus 0.5%. Every top-up gets a GST tax invoice (CGST+SGST or IGST) you can claim input-tax credit on; exports get a zero-rated invoice under LUT. Auto top-up through UPI AutoPay and e-mandates.
See pricing and feesHard daily and monthly budgets per key and project that pause instead of overspending, model allow-lists, concurrency caps, per-project cost attribution, audit-ready ledger exports and invoices you can pass through to clients.
How budgets workDeclarative routing policy with failover across hosts, price/latency/uptime scoring, canary and shadow. Published per-endpoint prices, our markup, 30-day uptime and TTFT so the cheapest host is never a surprise.
Browse modelsLogging on or off, retention 7–365 days, zero-data-retention mode that restricts routing to ZDR hosts, no-train filters, redaction before archive and a signed DPA. Processing regions you choose, including an India cell.
Read the DPA₹1,000 or $10 minimum by UPI, card, net banking or Stripe. Credits are USD; the rate is booked at purchase so usage carries no FX risk.
https://api.ai.ml/v1 for OpenAI-style clients, /v1/messages for Anthropic, /v1beta for Gemini. Same request bodies, same streaming.
Each response carries its cost lines. The ledger, budgets, request log and invoices live in the console; webhooks tell you when a budget trips.
Status, incidents and provider uptime are public.
Open the status page