Short answer, pay only for AI work, then match the plan to your pattern #
AI credits are BYOB's internal currency for AI powered operations and deployment. You buy credits once, then spend them when the AI generates code, edits code, processes vision or voice input, or publishes your site. Manual code editing, browsing, version control, viewing history, local testing, and exporting code do not consume credits.
Think of credits as the electricity meter and the subscription as the wiring. The wiring sets your capacity. The meter bills what you actually burn.
There are two ways to buy. Pay as you go top ups fit irregular or bursty work with no recurring commitment. Pro and Max subscriptions fit predictable monthly build volume with monthly included credits, improved top up economics, and paid feature gates. Purchased credits never expire. Free credits expire after 30 days.
This split mirrors how usage based billing works across software. Stripe documents metered billing models where customers pay for what they consume rather than flat seats (https://stripe.com/docs/billing), and its payments docs cover the top up and card flows that make bursty purchasing smooth (https://stripe.com/docs/payments). Paddle covers similar subscription plus consumption patterns for SaaS billing in its developer docs (https://developer.paddle.com/) and billing resources (https://paddle.com/resources). BYOB applies the same logic to AI work: flat plans for the platform, meters for the generation.
What do credits pay for and what stays free? #
Credits keep pricing tied to actual AI usage rather than seat time.
What consumes credits: AI code generation for pages, features, and components. AI refactoring and edits through chat. Vision processing from screenshots or designs. Voice commands. One click deployments.
What stays free: manual code editing in the editor. Browsing your project. Version control, snapshots, and history. Testing your site locally. Exporting code.
Credits also cover infrastructure usage such as hosting, CDN, and SSL. Deployed sites stay online as long as the account is active, with no ongoing credit drain for hosting. BYOB shows estimated credit cost before AI operations run, and usage history shows every transaction with what it was for. No mystery charges. You see the meter before it spins.
| Item | Credits feel | How to think | Note |
|---|---|---|---|
| Landing generation | Tens | One time build | Depends on complexity |
| Blog or section | Tens | One time build | Reuse lowers average |
| Small edits | Handful to tens | Iteration | Batch related prompts |
| Complex feature | Low hundreds | Multi step build | Clear spec cuts reruns |
| Deployment | Few | Per publish | Test locally first |
Try it: BYOB Credit Estimator
What do things actually cost in credits? #
BYOB credit costs follow practical bands from the linked BYOB pricing guides: landing-page generation and blog sections run tens of credits, small edits run a handful to tens, complex features run into the low hundreds, and each deployment runs a few credits. Scope: light monthly use lands in the hundreds of credits, active use around a thousand, and heavy agency use in the several thousands. Constraint: bands vary with complexity, so map your own pattern with the plan calculator guide before committing.
The BYOB source guides give practical bands, not exact quotes, because complexity varies. As described in the credit based pricing guide (https://byob.studio/blog/byob-credit-based-pricing-explained), a landing page generation runs tens of credits, a blog section runs tens of credits, a focused refactor runs tens of credits, small edits run a handful to tens of credits, complex features run into the low hundreds, and each deployment runs a few credits.
Translated to monthly patterns, light use with a few pages, a handful of edits, and several deploys runs hundreds of credits. Active use with more pages and dozens of edits runs around a thousand credits. Heavy agency use with dozens of pages, many edits, and frequent deploys runs several thousand. Most solo builders sit in the light to active range.
Regional pricing is auto detected with local currency and payment support for India alongside international card pricing. New accounts receive starter credits, enough to build simple pages, try iteration, and deploy several times. As described in the plan calculator guide (https://byob.studio/blog/byob-plan-calculator-credits), sliders help you map your own pattern to a tier before committing.
How does pay as you go compare to Pro and Max? #
Choose pay as you go in BYOB for spiky work with quiet weeks, Pro for steady solo or small-team production, and Max for high-throughput agency volume. Pro and Max add monthly included credits, improved top-up economics, and paid feature gates such as testing and domains. Constraint: tiers lose when the grant idles, so start monthly and move up only when consistent top-ups show under-tiering. Subscriptions exist because weekly shipping teams need forecastable spend and fewer mid sprint interruptions. As described in the subscriptions guide (https://byob.studio/blog/byob-subscriptions-pro-max-guide), the tiers break down by rhythm.
Pay as you go suits sporadic usage. Buy only when needed with no recurring commitment. Discovery phase startups with quiet weeks between launches often start here.
Pro suits steady solo founders, freelancers, and smaller teams with moderate weekly changes. Monthly included credits plus usually better top up economics plus access to paid gated areas such as testing and domains. Teams stop making micro decisions on every run.
Max suits agencies and high output teams running several active projects with daily iteration. Capacity and economics match throughput.
Enterprise suits governance, identity, white label, and support requirements rather than just more volume.
Moving from pay as you go to Pro changes workflow and billing. Illustrative examples (not customer cases): a freelancer with several active clients and regular edits is a typical Pro fit; an agency shipping many projects per month is a typical Max fit; a startup in discovery with quiet weeks between launches often starts pay as you go and moves to Pro after usage stabilizes.
How do you buy, track, and avoid the wrong tier? #
Buy and track BYOB credits in 3 steps: purchase through the Buy Credits flow, review usage monthly, and move up one tier when clean prompts still run out early. Prerequisites: try the free starter credits first to baseline one month of real use; the minimum purchase is 100 credits. Constraint: credits are non-refundable with no bulk discounts at standard volumes, so estimate from recent actual usage including reruns and deploys. Purchases run through the Buy Credits flow: choose an amount, pay by card or UPI where supported. The minimum purchase is 100 credits and credits appear instantly. There are no bulk discounts at standard volumes. Credits are non refundable, so trying the free starter credits first is the sensible move.
For tier choice, estimate from recent actual usage rather than best case months. Count feature builds, refactors, deploys, and testing overhead together, because failed prompt loops and frequent small deploys still consume credits. Volatile teams should add a larger buffer than stable teams. Higher tiers change both included capacity and effective top up economics, so consistent heavy top ups on a lower tier usually mean under tiering.
Run a short monthly review: credits consumed, top up frequency, reruns from unclear prompts, features shipped, and next month workload. If reruns are high and shipping is low, fix prompt quality before upgrading. If capacity runs out early despite clean prompts, move up one tier.
Which cost control habits matter more than tier? #
Detailed prompts cost about the same as vague prompts but succeed more often on the first try. Specific scope and acceptance criteria reduce iterations. Bundling related requests into one coherent instruction beats many fragmented requests. Manual edits suit tiny changes such as a button color. Testing locally before deploying avoids wasted deployment credits.
The break even framing is simple. Credits win for bursty usage across uneven months. Subscriptions win for constant daily generation. Prompt quality improvements often save more credits than tier changes, so tighten instructions before paying for more capacity.
Start on free credits, watch one month of real usage, then pick the wiring that fits the meter. The bill should follow the work, not lead it.
What are the trade-offs? #
Base tier for normal months plus top ups for spikes wins when patterns differ. Purchased credits never expire, free starter credits expire after 30 days, manual edits and version control stay free, and generation, edits, vision, voice, and deploys consume credits.
| Pick pay as you go when | Pick Pro or Max when |
|---|---|
| Work is spiky with quiet weeks between | Monthly volume is steady and predictable |
| Bursts need a few credits without a subscription | Top up economics on higher tiers beat one off buys |
| The site is mid experiment | Solo production or agency throughput fills the grant |
Tiers lose when the grant idles. Prompt quality saves more than plan shopping: scoped asks plus hand edits for copy beat regenerations. Pick the alternative the month the pattern flips.
Who this is for (and who should skip it) #
BYOB credit pricing fits builders mapping AI generation, edits, and deployment spend to a monthly rhythm, including beginners learning what burns credits, founders matching bursts against Pro or Max, and freelancers tracking client build costs. Skip subscription math until you have one month of real usage on free starter credits. Limit: failed prompt loops and frequent small deploys still consume credits, and bought credits are non-refundable.
This explainer helps builders mapping BYOB credit use to a monthly rhythm. If you ship regularly and you want to know when pay as you go stops fitting, the bands and review habit give you a tier to pick.
Skip subscription math until you have one month of real usage on free starter credits. Bursty months reward top ups, steady months reward Pro or Max for fewer mid sprint interruptions.
One limit to know. Failed prompt loops and frequent small deploys still consume credits, and bought credits are non refundable. Run one month on real use and fix brief quality before moving up a tier.
- Best for beginners learning what burns credits and what stays free in BYOB.
- Best for founders matching pay-as-you-go bursts against Pro or Max steady volume.
- Best for freelancers tracking client build costs without picking the wrong tier.
What we learned building this #
Credit pricing in BYOB is simple: Pro and Max plans bundle credits with monthly or annual billing, and the built-in credit estimator turns your project size into a plan suggestion. Every AI run shows an estimate before it executes, and the estimator compares pay as you go to subscription for steady versus spiky work.
