Dualyne documentation

Dualyne gives you every major AI model through one OpenAI-compatible endpoint. If your code already calls OpenAI, you only change two values.

Quickstart

1. Connect a wallet and create a key in the dashboard. 2. Set it as an environment variable. 3. Point your SDK at Dualyne.

export DUALYNE_KEY="dly_live_..."

import os
from openai import OpenAI
client = OpenAI(base_url="https://api.dualyne.com/v1", api_key=os.environ["DUALYNE_KEY"])
reply = client.chat.completions.create(
    model="claude-swift",
    messages=[{"role": "user", "content": "Hello"}],
)
print(reply.choices[0].message.content)

Authentication

Send your key as a bearer token on every request. Keys belong to the wallet that created them and stop working the moment you revoke them.

Authorization: Bearer dly_live_...

Never put a key in front-end code or a public repository. If a key leaks, revoke it in the dashboard and create a new one. Dualyne stores only a hash of each key, so a lost key can't be shown again.

Chat completions

POST/v1/chat/completions

Other OpenAI fields such as temperature, response_format and stream_options pass through unchanged. Tool calling and JSON mode work on models whose provider supports them. The response's model field names the exact model version that answered. n must be 1.

Models

GET/v1/models returns every model your key can use. Explorer wallets see fast models; Holder and Builder see the full catalog.

Each id is a stable name for a model family. When a newer version replaces the one listed under "Runs on", this table and the catalog update, and your code keeps working without changes.

Streaming

Set "stream": true to receive server-sent events. Each event carries a small piece of the answer; the last one is data: [DONE]. Lines starting with : are keep-alive comments and can be ignored (the OpenAI SDKs do this for you).

for chunk in client.chat.completions.create(model="gpt", messages=msgs, stream=True):
    print(chunk.choices[0].delta.content or "", end="")

curl -N https://api.dualyne.com/v1/chat/completions \
  -H "Authorization: Bearer $DUALYNE_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-swift", "stream": true,
       "messages": [{"role": "user", "content": "Hello"}]}'

To get token counts at the end of a stream, send "stream_options": {"include_usage": true}. The final chunk then carries usage and an empty choices list, so check the list before reading from it.

Limits and billing

Allowances reset at 00:00 UTC. Explorer and Holder responses include x-dualyne-remaining so your app can see how many requests are left today. A larger max_tokens than your tier allows is lowered to the cap. Each key can send up to 120 requests a minute.

Builder wallets top up with USDG or ETH from the dashboard. Each request reserves its worst-case cost (model price + 15%) and is then charged the real amount; the difference goes straight back to your balance. The dashboard lists every top-up and what each key spent.

Free tiers share a daily budget paid for by the treasury. On the rare day it runs out, free requests return 429 until 00:00 UTC; Builder requests keep working.

Errors

Errors use the OpenAI format, so SDKs raise them as normal exceptions:

{
  "error": {
    "message": "Daily allowance used up (20 requests for Explorer). It resets at 00:00 UTC.",
    "type": "rate_limit_error",
    "code": "quota_exceeded"
  }
}