OpenAI-compatible API across 4 providers

Quick start

1. Create a workspace — sign up and you get a <slug>.worouter.com subdomain with $1 of starter credit.
2. Mint an API key on the API Keys page. It looks like wor_live_….
3. Point any OpenAI SDK at your workspace's /v1 base with that key. Switch models by changing the model string.

Your API base URL

Video generation (async)

Video models (e.g. bytedance/seedance-1.5-pro) run as asynchronous tasks: create one, then poll until status is succeeded and download content.video_url. Billing uses the task's reported video tokens.

curl https://<your-workspace>.ai.wozinga.com/v1/videos/generations \
  -H "Authorization: Bearer wor_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "bytedance/seedance-1.5-pro",
    "content": [{"type":"text","text":"A sailboat at sunset --resolution 720p --duration 5"}]
  }'

Then poll with the returned task id:

curl https://<your-workspace>.ai.wozinga.com/v1/videos/generations/cgt-2026... \
  -H "Authorization: Bearer wor_live_..."

Example — curl

curl https://<your-workspace>.ai.wozinga.com/v1/chat/completions \
  -H "Authorization: Bearer wor_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-haiku-4.5",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Example — Python (openai SDK)

from openai import OpenAI

client = OpenAI(
    base_url="https://<your-workspace>.ai.wozinga.com/v1",
    api_key="wor_live_YOUR_KEY",
)

resp = client.chat.completions.create(
    model="anthropic/claude-haiku-4.5",
    messages=[{"role": "user", "content": "Hello!"}],
)

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

Example — Node (openai SDK)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://<your-workspace>.ai.wozinga.com/v1",
  apiKey: "wor_live_YOUR_KEY",
});

const resp = await client.chat.completions.create({
  model: "anthropic/claude-haiku-4.5",
  messages: [{ role: "user", content: "Hello!" }],
});

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

Streaming

Add "stream": true to any request and the response is a Server-Sent Events stream of chat.completion.chunk objects — same shape as OpenAI.

curl -N https://<your-workspace>.ai.wozinga.com/v1/chat/completions \
  -H "Authorization: Bearer wor_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-haiku-4.5",
    "stream": true,
    "messages": [{"role":"user","content":"Tell me a story."}]
  }'

Available models (31)

Routed IDs are provider/model. Sell price is per million tokens (USD): prompt / completion. Up-to-date list comes from the live catalog.

What's compatible?

• Chat completions: /v1/chat/completions (streaming + non-streaming)
• Model listing: /v1/models (returns the live catalog with pricing)
• Tool calls: OpenAI-shaped tools + tool_choice; translated to each provider's native form
• Usage metering: every response includes prompt_tokens / completion_tokens; the gateway debits your workspace's credit balance accordingly
• Per-key rate limits and insufficient-quota (402) responses