Principle01
Pricing maps to product actions
Users buy and spend credits for things they understand — chat replies, image generations, agent runs — not input/output token counts.
Why now
We're building the billing layer for AI products
AI products need a pricing surface users understand and a system you can meter, bill, and tune without exposing token math. Credit-first wallets, event metering, Stripe top-ups, SDK + MCP, and pricing intelligence — for teams shipping real AI into production.
What ships in the product
- Wallets
- Event metering
- Stripe top-ups
- SDK + MCP
- Pricing Advisor
The current model breaks here
AI monetization is still awkward
One extreme
Flat pricing guesswork
Fixed plans that don't map to usage. Margins unknown.
Between themChargly
Credit-first layer: actions users get, economics you control.
Other extreme
Raw token complexity
Expose provider math. Users don't want token counts.
Chargly gives teams a credit-first billing layer — maps AI usage to product actions, keeps token chaos out of sight, scales from prototype to production.
Operating model
How Chargly works
One flow from purchase to metering to pricing — the shape of the product, not a feature list.
Our point of view
What we think should be true
Non-negotiables for how credit billing, metering, and pricing changes should work in AI products.
If AI billing is going to work well, these are the principles we think should hold.
Principle02
Credits beat raw tokens
Credits are a better product-facing abstraction. They hide provider complexity and let developers price AI actions clearly.
Principle03
Developers need control and visibility
You decide what each action costs. You see usage, margins, and recommendations. Nothing changes without your approval.
Principle04
Pricing changes are explainable and versioned
Recommendations come with reasons. Every change creates an immutable version. No black-box pricing.
What Chargly ships today
The system we're building
One stack: wallets, metering, checkout, developer surfaces, and advisor-style pricing — wired together, not bolted on.
Balances, metering, checkout
Where money moves
- 1
Credit wallets
Per-user balances. Deduct on events. Top up via Stripe.
- 2
Event metering
Define actions, set credit costs. Meter every AI call.
- 3
Stripe-powered top-ups
Checkout flows, webhooks, automatic balance sync.
Surfaces & pricing intelligence
How you integrate and tune
- 4
SDK + MCP support
npm package and MCP server for agent-native workflows.
- 5
Pricing Advisor
Recommendations, version history, apply/reject controls.
What this becomes
From billing infrastructure to pricing intelligence
1Today
Billing infrastructure
Credits, wallets, metering, top-ups.
2Next
Pricing intelligence
Recommendations, version history, apply/reject controls.
3Later
Operational layer
Cost visibility, margin control, and workflows that scale.
For teams shipping AI
Building with Chargly? Let's talk
If you're wiring credits, wallets, and metering into a real product, we're happy to compare notes — docs, demo, or a direct line to the team.