v1.0 — Now in Beta

One API. Every AI Model.

OpenAI-compatible endpoint. Credit-based billing. Smart routing across providers.

16+
models available
<200ms
average latency
99.9%
uptime

Start in 30 seconds

bash
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
# Install the SDK
"docs-token-command">npm install @conflux/router

# Or use any OpenAI-compatible client
"docs-token-command">curl https://theconflux.com/v1/chat/completions \
  -H "Authorization: Bearer cf_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "task_type": "code_gen",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

# Or specify a model directly
"docs-token-command">curl https://theconflux.com/v1/chat/completions \
  -H "Authorization: Bearer cf_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

It's just the OpenAI API. Swap the base URL and key.

Why developers switch

🔌

Drop-in Replacement

Works with the OpenAI SDK, LangChain, Vercel AI SDK, and any OpenAI-compatible client. Change 2 lines of code.

💰

Transparent Pricing

Pay only for what you use. Credits never expire. No monthly minimums. See exact costs before you call.

⚡

Smart Routing

Automatic failover between providers. Route by cost, speed, or quality. Your app never goes down.

Ready to build?

Free tier includes 500 credits. No credit card required.