Simple, metered, prepaid.
No subscription. No revenue share. You pay a small platform fee per unit of usage we serve for you — typically around 1% of what your customers pay you. Everything else (storefront, sign-in, credit billing, token metering) is included.
Platform fees
| Usage | Fee | In practice |
|---|---|---|
| Text request (any provider) | $0.00005 / request | $0.05 per 1,000 requests |
| Image generation | $0.0001 / image | $0.10 per 1,000 images |
| Realtime audio (voice, speech-to-text, text-to-speech) | ~$0.001 / minute | $0.06 per hour |
| Realtime video | ~$0.001 / minute | $0.06 per hour |
Same rates across Gemini, OpenAI, Claude and Soniox. Fees are deducted from your prepaid balance as usage happens — your balance is your spending cap, so there are never surprise bills.
Getting started
$0.50 free credit
Every new developer account starts with $0.50 of platform credit — about 10,000 text requests or 8 hours of realtime audio. Enough to build, launch and test for real.
Top up from $5
When you're ready for customers, top up your balance via PayPal — any amount from $5 to $1,000. Your credit doesn't expire.
Hard stop at $0
Serving pauses when your balance runs out and resumes the moment you top up. A low-balance email warns you first, at a threshold you set.
A worked example
Say you sell a voice-app credit pack: your customer pays $7 for 60 minutes of talk time, straight into your PayPal. If they use every minute, our fee is 60 × $0.001 ≈ $0.06 — about 0.9% of your revenue. Your AI provider bills you separately for the model usage (you bring your own keys); Lumen never sits between you and either of them.
Your dashboard's pricing page shows this calculation live for every offer you create — revenue, platform fee at full consumption, and net.
What the fee covers
- Your hosted storefront at
your-slug.lumencredits.com - End-user sign-in and API keys for your customers
- Credit packs and subscriptions sold through your PayPal — we never touch customer money
- Per-token, per-model metering across Gemini, OpenAI, Claude and Soniox
- Realtime voice infrastructure (WebSocket bridging, Opus transcoding)
Compare: billing-only platforms typically take 1–5% of your revenue plus payment fees — and you still have to build the metering and host the app.