Getting started

Sign in at app.router4ai.app. Go to API Keys → Create key. You'll see the full key once — copy it immediately.

ram_live_yourkeyhere...

Keys start with ram_live_ and are hashed (Argon2) on our side. Lost keys can be rotated from the dashboard.

Works with any OpenAI-compatible SDK or HTTP client. Only the baseURL changes:

Node / TypeScript

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.RAM_API_KEY,              // ram_live_xxx
  baseURL: 'https://api.router4ai.app/v1',
});

const response = await client.chat.completions.create({
  model: 'gpt-4o-mini',
  messages: [{ role: 'user', content: 'Hello' }],
});

console.log(response.choices[0].message.content);

Python

from openai import OpenAI

client = OpenAI(
    api_key="ram_live_xxx",
    base_url="https://api.router4ai.app/v1",
)

response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello"}],
)

print(response.choices[0].message.content)

curl

curl https://api.router4ai.app/v1/chat/completions \
  -H "Authorization: Bearer ram_live_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role":"user","content":"Hello"}]
  }'

Models are auto-routed to one of three tiers. You can pass any of these model IDs and the router picks the best provider:

You can also pin a specific provider by prefixing the model:openai:gpt-4o-mini, anthropic:claude-sonnet-4-6, etc.

Every response includes a _ram field so you can audit routing decisions:

{
  "id": "chatcmpl-...",
  "choices": [...],
  "usage": { "prompt_tokens": 14, "completion_tokens": 18 },
  "_ram": {
    "tier": "fast",
    "provider_used": "openai",
    "failover_count": 0,
    "latency_ms": 527,
    "cost_billed_cents_usd": 0.0028,
    "cache_hit": false,
    "request_id": "84f9a90c-..."
  }
}

You start on a free trial: 1,000,000 tokens valid for 14 days. After the trial (or whenever you want), upgrade via Billing to Pro ($49/mo) or Business ($499/mo). Stripe charges your card once per billing period with a single invoice covering base fee + usage overage.

• • Cancel anytime — your keys keep working until the period ends.
• • Usage is reported to Stripe Meter Events per request.
• • Invoices, receipts, and card updates live in Stripe Customer Portal.

Default: 60 requests/second per tenant. Need more? Contact support. Limits are enforced via Upstash Redis, so bursts over short windows are tolerated.

Email support@router4ai.app. For security issues, use the subject “SECURITY”.