One API key.
Every AI model.
Lumen 用一把金鑰，把你連接到 248+ 個 AI 模型。用它來建構應用、執行代理，或者跳過設定，直接在本站與角色聊天。同一把金鑰，同樣的速度，同樣的價格。
from openai import OpenAI
client = OpenAI(
base_url="https://lumen-api.develse.com/v1",
api_key="YOUR_API_KEY",
)
reply = client.chat.completions.create(
model="claude-haiku-5-5",
messages=[{"role": "user", "content": "Hello!"}],
)Works with any OpenAI SDK. Just change the base URL.
Built for production.
Everything you need to ship AI features without juggling providers, keys and invoices.
Drop-in compatible
OpenAI, Anthropic and Gemini formats. Point your SDK at Lumen and keep your code.
Automatic failover
When a provider slows down or fails, requests move to a healthy one. No retry logic to write.
Pay as you go
Clear per-token prices, free models to start, and credit that never expires.
Keys with limits
Create a key per project and cap it with model allow-lists, spending limits and expiry dates.
Usage analytics
Spend, latency and success rate per model and per key, for any date range.
Private chat
The built-in chat keeps your conversations in your browser, not on our servers.
選擇你真正使用的用戶端。
Lumen 可作為自訂的 OpenAI 相容 API，在程式設計代理、角色聊天用戶端、AI 聊天用戶端與 CLI 工具之間通用。
報到，新兵。收好裝備，十秒後躍遷。