Unified AI API · 99.98% uptime

One key. Every frontier model.

Claude, GPT, Gemini, Grok and DeepSeek through a single OpenAI-compatible endpoint. Transparent billing. Automatic failover. Global payments.

Get your API key See pricing

43ms median TTFT 180+ models Stripe · Airwallex · PayPal

Illustrative demo metrics

latiao — zsh LIVE

$ curl -s https://api.latiao.ai/v1/chat/completions \

-H "Authorization: Bearer $LATIAO_KEY" \

-d '{"model":"claude-sonnet-4.5","stream":true,

"messages":[{"role":"user","content":"Ship it."}]}'

HTTP/2 200 · 43ms ttft · cache hit 68%

data: {"delta":{"content":"Shipping"}}

data: {"delta":{"content":" is a"}}

data: {"delta":{"content":" decision."}}

data: {"usage":{"in":12,"out":3,"cost_usd":0.00042}}

✓ routed · 1 attempt · $0.00042

12,438req / min

0.998success

$0.00042per 1k

Models

claude-opus-5.5 claude-sonnet-4.5 claude-haiku-4.5 gpt-6-astra gpt-6-mini gpt-5.5-codex gemini-3.5-pro gemini-3.5-flash gemini-3-ultra grok-5 deepseek-v4.1 deepseek-r2 qwen-4-max kimi-k3 mistral-large-3 llama-5-405b qwen-4-vl whisper-3

01 — Why LATIAO

Built for people who ship.

One endpoint, health-aware routing, and billing you can actually audit.

/ 01

One endpoint

Drop-in OpenAI, Anthropic and Gemini compatibility. Change the base URL, keep your code. Streaming, tools and vision are supported on every path.

/ 02

Always on

Health-aware scheduling and automatic failover across accounts, channels and regions. A bad upstream never becomes your outage.

/ 03

Every token, explained

Per-request cost, cache hits, reasoning tokens, TTFT and attempts. Export CSV or reconcile against your own ledger.

02 — Developers first

From curl to production in 3 lines.

Keep your SDK. Change two values. Works with Claude Code, Codex, Cursor, OpenCode, Cline, Cherry Studio and any OpenAI-compatible client.

OpenAI compatible Anthropic /v1/messages Gemini v1beta

Read the docs Client setup

# one endpoint, any model
curl https://api.latiao.ai/v1/chat/completions \
  -H "Authorization: Bearer $LATIAO_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4.5",
    "messages": [{"role":"user","content":"Ship it."}],
    "stream": true
  }'

from openai import OpenAI

client = OpenAI(
    base_url="https://api.latiao.ai/v1",
    api_key="lt-...",
)

stream = client.chat.completions.create(
    model="claude-sonnet-4.5",
    messages=[{"role": "user", "content": "Ship it."}],
    stream=True,
)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.latiao.ai/v1",
  apiKey: process.env.LATIAO_KEY,
});

const stream = await client.chat.completions.create({
  model: "claude-sonnet-4.5",
  messages: [{ role: "user", content: "Ship it." }],
  stream: true,
});

03 — Pricing

Pay for what you use. Nothing else.

No seats. No hidden fees. Top up in your currency; every request is itemized.

Demo data Illustrative plans for design review — not a live quote. See the pricing page for the full table.

Pay as you go

Starter

$0 / month

No subscription. Top up in USD and pay per request.

180+ models on one endpoint
Per-request itemized billing
Automatic failover and health-aware routing
Usage export (CSV)

Start free

Pro

$20 / month

For individuals shipping production traffic every day.

$20 included credit each month
5% lower model rates
Priority routing during peak load
Higher rate limits per key

Go Pro

Teams

Team

$99 / month

Shared billing, invoices and a data processing agreement.

$100 included credit each month
Up to 5 members with per-member quotas
10% lower model rates
Invoices, VAT ID and DPA

Talk to us

Compare plans and model prices →

SOC 2 roadmap GDPR ready Stripe & Airwallex No training on your data 99.9% SLA target Upstream details never exposed

99.98%

30-day uptime

43ms

median TTFT

180+

models

12,438

requests / min

Demo data Illustrative figures, not live telemetry.

04 — FAQ

Questions, answered.

Everything engineers ask before moving production traffic.

Which models can I call with one key?

Claude, GPT, Gemini, Grok, DeepSeek, Qwen, Kimi, Mistral and Llama families through one OpenAI-compatible endpoint. Model names and availability are listed on the models page.

Do I need to change my code?

No. Point your existing OpenAI SDK at https://api.latiao.ai/v1 and swap the API key. Anthropic and Gemini native protocols are also supported on their own paths.

How does billing work?

Every request is itemized by input, output, cached and reasoning tokens. You prepay a USD balance and are charged per request. Subscriptions add included credit and lower rates.

What happens when a model is unavailable?

Health-aware routing detects failures per account, channel and region, then retries the next healthy option. A bad upstream never becomes your outage, and the attempt chain is visible in your request log.

Are my prompts used for training?

No. Prompts and completions are not used to train models. Request bodies are not written to application logs; metadata is retained for billing and abuse prevention only.

Which payment methods are supported?

Card payments through Stripe and Airwallex, plus PayPal. Multi-currency display is available; the ledger settles in USD so invoices reconcile exactly.

Can I export usage data?

Yes. Usage, request logs and invoices export to CSV from the console, and the API exposes the same aggregates for your own dashboards.

Is there an SLA?

Paid plans target 99.9% monthly availability. The exact credit terms are defined in the service agreement and are pending legal review for this preview.

Ready to ignite?

Start with $1.
Get 180+ models.

Create an account, top up in your currency, and send your first request in under five minutes.

Create account Read the docs

No card required to start Cancel anytime GDPR ready

One key. Every frontier model. Claude. GPT. Gemini. Grok. DeepSeek.