Quickstart
Lumen API is OpenAI-compatible. If you can call OpenAI, you can call Lumen — change the base URL and the model.
1. Point your SDK at Lumen
# Python — openai>=1.0
from openai import OpenAI
client = OpenAI(
base_url="https://lumen-api.eliteaiempire.com/v1",
api_key="lk_live_YOUR_KEY",
)
2. Let the router decide
resp = client.chat.completions.create(
model="lumen-auto",
messages=[{"role":"user","content":"Summarize this in one line: ..."}],
)
print(resp.model) # e.g. "groq/llama-3.3-70b" — the model Lumen chose
print(resp.usage.cost_usd) # what it cost
print(resp.usage.saved_usd) # vs calling frontier directly
3. Or pin a tier / set a quality floor
# Force a minimum quality class — Lumen still picks the cheapest model that clears it
resp = client.chat.completions.create(
model="lumen-auto",
messages=[...],
extra_body={"lumen": {"quality_floor": "frontier"}},
)
# Or pin a specific upstream when you must: model="anthropic/claude-..."
Model names
lumen-auto (recommended), or pin provider/model across our 8 vetted providers. Banned providers (CN/RU) are never available, even by pin.
Streaming, tools, vision
SSE streaming, function/tool calling, and vision pass through to the chosen upstream. The router accounts for tool-use and long-context when scoring difficulty.
Budgets & keys
Per-key monthly budget caps, hard/soft limits, and rotation. Indie gets 5 keys, Pro+ unlimited. Every call lands in the audit log.
Full reference docs ship with the public API launch. Want a key now? Join the early-access waitlist — founding developers get 20% off forever and higher free limits.