8/21 demo: Building component libraries from Figma with AI

What are best AI tools? Take the State of AI survey

Builder.io
Contact sales
Builder logo
builder.io

Blog

Home

Resources

Blog

Forum

Github

Login

Signup

×

Visual CMS

Drag-and-drop visual editor and headless CMS for any tech stack

Theme Studio for Shopify

Build and optimize your Shopify-hosted storefront, no coding required

Resources

Blog

Get StartedLogin

☰

‹ Back to blog

AI

Introducing Usage-Based Agent Credits

August 7, 2025

Written By Josh Adams

Starting August 14, 2025, AI Credits will follow a usage-based consumption model we call Agent Credits.

Instead of paying per message, you consume credits based on the tokens the agent uses while it plans, reasons, calls tools, and generates code. Learn more in the full Agent Credits documentation.

What's changing:

  • Usage-based pricing: Credits for what you use instead of per-message limits
  • Rollover with cap: leftover credits carry forward until your balance reaches 2× your monthly amount; additional credits do not accumulate beyond that cap.
  • Unified currency: One credit system for all AI operations

How agentic AI delivers better results:

  • Adaptive reasoning: The AI uses tokens as needed to understand your context and provide the best solution
  • Quality over efficiency: Instead of rushing to a quick answer, the system considers approaches and edge cases
  • Intelligent resource allocation: Complex tasks receive more computational power, while simple requests use less.
  • No artificial limitations: The AI isn't constrained by arbitrary usage caps that would compromise quality.

Why Agentic AI needs usage-based pricing

Our agent thinks and works like a teammate, not a chatbot. Simple prompts finish quickly, while complex challenges trigger deeper exploration and tool calls. These sessions naturally consume more tokens, but deliver higher-quality results.

Usage-based pricing ensures:

  • Quality first: The agent can think as long as needed without hitting arbitrary message caps.
  • Fair costs: You pay only for the computation you actually use.
  • No hidden limits: When a task gets hard, the agent scales up automatically.

Per-message pricing assumes every request is the same size. In practice, an agent might tweak a line of code or architect an entire feature. Token usage reflects that difference, so Agent Credits charge exactly for the work your agent performs.

Along with these changes, we're adding rollover functionality. Unused credits now carry forward, so teams have flexibility across project cycles.

Agent Credits provide complete cost transparency with a straightforward formula:

Credit Cost = Pass Through LLM Token Cost × (1 + 0.25 margin)

Base Credit Value: $0.05

  • Model cost: $0.04
  • Our margin: $0.01 (25%)

You know exactly what you're paying for: the actual model cost plus our service fee. No hidden markups or surprise charges.

Wondering how these credits translate to everyday tasks? Check out the real-world examples in our Agent Credits documentation.

All customers: Automatic transition on August 14th

How does this change impact my current plan?

You can view Agent Credit limits by product and plan in our docs here.

Will my monthly bill change?

Credit allocations are changing as we move to token-based pricing. Your actual usage costs will depend on the complexity and length of your AI requests.

How do I purchase more credits?

Pro plans: You can purchase additional credits in-app.

Enterprise plans: contact sales team to learn more enterprise options.

What happens if I run out of credits?

You'll see clear status indicators in your dashboard. New AI operations pause until you purchase more credits or wait for the next billing cycle.

How does rollover work?

Pro and Enterprise plans roll over unused credits up to your monthly allocation (maximum 2× your monthly amount). Free plans reset daily and monthly.

Agent Credits are available today for all new users. Existing users transition automatically.

Enterprise pricing ? Contact our sales team

Generate clean code using your components & design tokens
Try FusionGet a demo

Share

Twitter / X
LinkedIn
Facebook
Share this blog
Copy icon
Twitter "X" icon
LinkedIn icon
Facebook icon

Visually edit your codebase with AI

Using simple prompts or Figma-like controls.

Try it nowGet a demo

Design to Code Automation

A pragmatic guide for engineering leaders and development teams

Access Now

Continue Reading
AI4 MIN
Vibe code immersive 3D effects in one prompt
August 10, 2025
AI4 MIN
Convert HTML to Design in Figma
August 6, 2025
AI6 MIN
How to set up and use the Linear MCP server
August 5, 2025