Question 1

How fast can I integrate AdCrier into my AI app?

Accepted Answer

Five minutes for the SDK call. Your first commission can land within hours of going live, depending on your traffic.

Question 2

Will my users notice the affiliate cards?

Accepted Answer

Only as a helpful suggestion card. We don't modify your AI's text or inject sponsored language into your model. The card sits outside the response body and is fully under your visual control.

Question 3

Does AdCrier slow down my AI app?

Accepted Answer

Total overhead is 300–600ms — typically invisible because users are still reading the AI's response. Run it in parallel for zero perceived latency.

Question 4

What about GDPR and user privacy?

Accepted Answer

We don't store what your users say. Every userMessage and aiResponse passes through our classifier in memory and is discarded immediately — never written to disk, logs, or any persistent store. No cookies, no fingerprinting, no cross-site tracking.

Question 5

Do I need advertiser relationships to use AdCrier?

Accepted Answer

No. AdCrier is a two-sided marketplace — advertisers sign up directly, top up a balance, and bid CPC on the IAB categories your AI app covers. You never source advertisers yourself; we never insert a third-party affiliate network between you and them.

Question 6

How fast do I get paid?

Accepted Answer

Commissions confirm 30–60 days post-sale (network refund windows). Once received and reviewed, withdraw via Stripe Connect when your balance hits $500. No auto-payouts, no surprises.

Question 7

How much can I earn with AdCrier?

Accepted Answer

Earnings depend on traffic and intent mix. Typical AI-app eCPM ranges from $3 to $15, and you keep 85% of every commission. Use the on-page calculator to estimate based on your messages-per-day.

Question 8

Which LLMs does AdCrier work with?

Accepted Answer

AdCrier is model-agnostic. It works with OpenAI (GPT-4, GPT-4o, GPT-5), Anthropic (Claude), Google Gemini, Mistral, Meta Llama, Cohere, DeepSeek, and any other LLM — because the SDK only sees the message + answer text, not the model that produced it.