Credit-first billing
Charge users for product actions, not token math
Users buy credits for chat replies, image generations, agent runs. A stable pricing surface that maps to value, not provider economics. Wallets, metering, Stripe top-ups — one layer.
The challenge
AI usage is hard to productize
Raw token economics don't translate to product value. Users don't think in input/output tokens. They think in actions — replies, generations, summaries. Exposing provider costs creates confusion, support burden, and unpredictable pricing.
Credits vs tokens
Token-based billing creates friction
Provider costs fluctuate. Exposing raw tokens means users see pricing move for reasons they can't control. You need a stable abstraction that hides provider complexity while giving you margin control.
1 chat reply = 4 credits
User sees a product action
847 input + 312 output
User sees provider economics
Example
A chat request, credit-style
User sends a message. Your app calls the LLM, gets a reply. Instead of counting tokens and calculating cost, you meter one event: chat.reply = 4 credits. Wallet deducts 4. User sees "4 credits used." You control what 4 credits costs you (model, context) and what it costs them (your margin).
// User sends message → LLM reply
meter_event("chat.reply", userId) // 4 credits
How Chargly handles it
Credit wallets, event metering, Stripe top-ups
Define what each AI action costs in credits. Users buy credits, spend them on actions they understand, top up when low. Wallets hold balances. Metering deducts on events. Stripe handles checkout. You see margins and recommendations; users see a clean product experience. No token math, no provider exposure.
Try it
See how cost turns into credits
Turn provider cost into a clean credit price your users can understand. Adjust the assumptions and see how the pricing outcome changes live.
This is a simplified preview. Chargly handles the real pricing logic in production.
See how Chargly applies this in productionSuggested action price
4 credits
Users see credits. You keep control of margin underneath.
This pricing is strong, but may feel expensive unless the action has high perceived value.
Where it fits in the product
Wallets, metering, SDK, Pricing Advisor
Credit billing sits at the center. The SDK lets you meter events and deduct from wallets. Stripe powers top-ups. MCP exposes the same system to agent workflows. Pricing Advisor helps you adjust credit costs as provider economics change.
Add credit billing to your AI product
Wallets, metering, Stripe top-ups. Free to start.