Cost‑engineering an "AI Generate" button in a freemium product (from $0.08 to $0.029 per click)

Описание и применение

Оптимизация стоимости генерации визуализаций в freemium‑приложении снизила цену с $0.08 до $0.029 за клик, устранив расходы для бесплатных пользователей и повысив маржу платных подписчиков до 96%.

Задача

A freemium coding‑prep app’s AI‑generated visualizations cost ~$0.08 per click, making free‑tier usage unsustainable and eroding margins for paid users.

Результат

Reduced per‑click AI cost from $0.08 to $0.029, eliminating free‑tier cost and increasing Pro user gross margin from ~90% to ~96%.

Рабочий процесс

• 1. Detect user tier; if free, serve cached visualization or preview without calling Claude.
• 2. For Pro users, check cache; if miss, call Claude with cached system prompt and problem description.
• 3. Use prompt caching to reduce input token cost to ~$0.00084 per hit.
• 4. Apply output constraints and max_tokens=3500 to limit output to ~1,850 tokens, saving ~$0.032 per call.
• 5. For regenerations 1–5, route to Groq Llama 3.3 70B; beyond that, rate‑limit.
• 6. Run a Haiku classification pass offline to filter out non‑visualization‑worthy problems, saving ~13% of calls.
• 7. Store successful visualizations in DB for future cache hits.

Доказательства

At naive implementation, each click cost me about $0.08. Workable for Pro subscribers. Ruinous if I let free‑tier users click freely.

Источник

Through five cost decisions … I got the per‑click cost to $0.029 and pushed free‑tier users to a $0‑marginal‑cost path entirely.

Источник

Free users never trigger Claude.

Источник

Prompt caching charges ~10% of full input‑token price on cache hits.

Источник

Output tokens cost 5x more than input tokens on Sonnet 4.5.

Источник