Klip Kanvas Logo
Klip Kanvas
Create
FAQCreditsPricingVideo Generation

Credits and Rendering FAQ: What Costs Credits, Rollover, Queues and Failures

How Klip Kanvas credits work — what a render costs, rollover and expiry, failed jobs, queue speed, how to estimate a test batch, and how to buy more without guessing.

Updated 2026-07-0715 min read
EnglishEspañolItalianoNorskDanskÍslenska

Credits are how renders get paid for, not a second currency for every click. This FAQ covers how they work, what actually spends them, rollover, failed jobs, queue speed, how to estimate a batch, and how to buy more without treating a blog post as a price list.

01

How Credits Work

1.What is a credit?

A Klip Kanvas credit is the unit that pays for a successful video render. Editing a script, picking an avatar, previewing a timeline and saving a project do not spend it. The spend happens when you generate the actual file. That split is why you should do your thinking in the editor and your spending in batches. Free accounts start with 50 credits and no credit card. Paid plans include a monthly allotment. Exact pack sizes and prices change — use /pricing, not a screenshot from a review post. Credits are not cash and they are not seats.

#How Credits Work

2.Do I need a card to try the product?

No. The free plan is 50 credits with no credit card required, which is enough to render a real first comparison rather than a single toy clip. Use it for a structured test: more than one hook, more than one avatar, one body script. If you spend the 50 on twenty almost-identical 15-second clips, you will have learned nothing and you will think the product “doesn’t work.” The free allotment is an evaluation budget. Treat it like one. Upgrading is optional until you want volume, 4K, cloning or team features — those gates live on /pricing.

#How Credits Work

3.Is a credit the same on every plan?

The unit is the same; the allotment, rollover and extras around it are not. A 30-second 1080p render does not secretly cost more because you are on a higher tier. What higher tiers change is how many you get, how fast they render, whether 4K is available, and seats. Do not assume a “pro credit” is a different video. Assume a larger bucket and fewer gates. If a comparison site lists a credit as a dollar amount, ignore it and open /pricing. We will not freeze a unit price in this article.

#How Credits Work

4.When are credits deducted — at click or at completion?

They are committed when a render job starts and settled against a successful output. A job that fails in our pipeline is not supposed to eat the credit; a job you hate but that completed is. That distinction matters. Cancelling because you spotted a typo after the file is done does not refund. Watch the preview, then render. If a job is stuck without producing a file, use the failed-render path rather than starting three more of the same. Double-clicks are how people burn a free 50 without noticing.

#How Credits Work#Failed Renders
02

What Costs Credits

5.What actually costs credits?

Rendering a video. In Klip Kanvas, cost scales with length and output resolution, not with which avatar, voice or template you picked. A 15-second 1080p costs less than a 60-second 1080p, and 4K costs more than 1080p at the same length. Localising into another language is a new render, because voice and lip-sync are rebuilt. Re-exporting after a caption fix is a new render. Preview, script generation, hook lists and brand-kit edits do not. If you are unsure, look at the confirm step before you click — it is the last cheap place to catch a 3-minute 4K you did not mean to order.

#What Costs Credits

6.Does picking a different avatar or voice change the cost?

No. Casting tests should never be a budget decision. Three avatars on the same 30-second script cost like three renders of that script. That is why the standard first batch is 6 creatives — 3 hooks × 2 avatars — not one avatar forever. Voice clones and stock voices bill the same at render time; cloning itself may be plan-gated, which is a feature gate, not a per-minute surcharge. Do not skip a second face to “save credits.” Skip a fourth near-duplicate hook instead.

#What Costs Credits

7.Does 4K cost more than 1080p?

Yes. 4K is heavier to render and is a premium export. For Meta and TikTok, 1080p is the working master. Spend 4K on the cuts that need a client-facing or retail master, not on every hook test. If your plan does not include 4K, the option will not be a bargain button — it will be gated. Check /pricing. Rendering a whole test batch in 4K “to be safe” is one of the faster ways to empty a month’s allotment for no media-buying gain.

#What Costs Credits

8.Do previews, downloads or extra aspect ratios cost extra?

Preview in the editor does not. Each exported file does. 9:16, 1:1 and 16:9 are separate renders if you generate three files. Download of an already rendered file does not charge again. Team members downloading the same output do not each pay. The practical habit: preview, pick, render the ratio you will upload, add other ratios only for placements you buy. Agencies that “export everything in every size just in case” treat credits like office snacks. They are not.

#What Costs Credits
03

Rollover & Expiry

9.Do unused credits roll over?

On paid plans, unused credits roll over for one billing cycle. They do not stack forever. A light month plus a heavy month works; three light months will not build a war chest. Free-plan credits are for evaluation and should not be treated as a bank. If you are pausing spend, still watch the cycle date. Rollover is a cushion for campaign timing, not a savings account. When in doubt, read the current wording on /pricing and in-app, because promotional credits can have different clocks.

#Rollover & Expiry

10.What happens to credits if I upgrade or downgrade?

The new allotment follows the new plan at the next cycle or as described at checkout — do not assume unused credits convert at a secret rate. Downgrading does not usually increase the time you have to spend what you already have; it can change rollover and feature gates (4K, seats, cloning) immediately or at cycle end. Upgrade if you are blocked on a gate you need this week. Do not upgrade only to hoard credits. Check /pricing and the billing screen for the actual proration behaviour rather than a rule of thumb in an FAQ.

#Rollover & Expiry

11.Do credits expire if I cancel?

When you cancel, you generally keep access until the end of the period you paid for, and unused credits do not become a cash refund. After access ends, do not expect a bank of credits to wait indefinitely. Download the masters you need before you leave. This is a real limitation: credits are for rendering on the platform, not a store-of-value. If a campaign will run past cancellation, export first. Billing details and exceptions live in the refunds FAQ and at /pricing, not in a sentence that pretends to be a contract.

#Rollover & Expiry

12.Can I gift or transfer credits to another workspace?

No. Credits sit on the billing account they were granted to. Agencies should bill the workspace that should own the spend, or carry a house account they are willing to spend from. Transferring between client workspaces would mix ledgers we will not mix. If a client is leaving, export their videos; do not plan on sliding the remaining credits after them. Seat changes and invoices are a billing problem. Credits are a render problem. Keep those ledgers separate in your own ops too.

#Rollover & Expiry
04

Failed Renders

13.If a render fails, do I lose the credits?

A failure in our pipeline — no file, error state, job death — should not keep the credit. A completed file you do not like still spends. If the UI shows failed, wait for the refund to land before you hammer retry, then retry once. If it fails again, stop and contact support with the job id rather than burning the rest of the bucket. We cannot honestly say that every edge-case timeout has always refunded in milliseconds; if a credit is still missing after a confirmed failure, that is a support ticket, not a feature.

#Failed Renders

14.The video rendered but looks wrong — is that a “failed” render?

No. Wrong avatar energy, a typo, a bad crop or a claim you recanted is a successful render of a bad brief. Fix the project and pay for a new one. This is the same as a photo studio: they do not refund because you dislike the shirt. Use preview. For quality issues that are clearly engine bugs — a crashed frame, missing audio, a truncated file — treat those as failures even if a file downloaded. Attach the output and the job id. “I would shoot it differently” is not a bug.

#Failed Renders

15.A job is stuck at a high percent for a long time. What should I do?

Wait through the normal 3–5 minutes for a 30-second ad, longer for 4K or peak queue. Past that, mark it stuck rather than opening five duplicate jobs. Duplicates can all succeed later and all charge. Refresh once, check status, then support if it is still frozen. Do not delete the project while a job is in flight. Stuck is a queue or worker problem; we refund or complete those, but we cannot promise a specific minute on a bad day. If it happens in a launch window, that is why you do not render the whole catalogue at T-minus ten.

#Failed Renders#Queue & Speed

16.Can I retry a failed job without being charged twice?

Retry the same failed job through the retry control so the system can settle the first attempt first. Starting a brand-new render of the same project while the first is in an unknown state is how double spend happens. If the first later succeeds, you will have two files and two charges. That is avoidable ops, not a mystery. One retry, then a human. The UI is not a slot machine. Credits on failure is the policy; race conditions from panic clicks are on the clicker.

#Failed Renders
05

Queue & Speed

17.How long does a render take?

About 3–5 minutes for a typical 30-second ad on Klip Kanvas, and a first ad from sign-up still targets under 10 minutes including setup. Longer videos, 4K, localisation and busy periods add time. That range is typical, not a contract. If you need 40 files by morning, start the night before and stagger, rather than pasting 40 jobs into a peak-hour queue. Speed is good enough for same-day testing. It is not a live broadcast encoder. Plan media buying around “this afternoon,” not “this minute.”

#Queue & Speed

18.Is there a priority queue?

Higher tiers can include faster or priority rendering. Exact cutoffs sit on /pricing because they change with capacity. Free and entry jobs are real jobs, not a pretend queue, but they will wait behind priority when the farm is hot. If your workflow is “client in the room, need a file in five minutes,” do not put that demo on a crowded free account at peak time. Render a sample before the meeting. Priority is a plan feature, not something support can magically apply to one lucky job.

#Queue & Speed

19.Can I render a lot of ads at once?

Yes, within plan and fair-use limits. Batch is a normal workflow: name clearly, estimate credits first, then queue. QA a sample before you queue 100 near-duplicates. A bulk send of 80 almost-identical hooks wastes credits and produces unreadable tests. The product will let you spend; it will not make the test design smart. If the queue slows as you pile on, that is capacity, not a hidden error. Stagger huge jobs. Details for catalogue-scale work also live in the bulk-generation hub when you need ops beyond credits.

#Queue & Speed

20.Why is my render slower than 3–5 minutes today?

Length, 4K, extra language passes, and how many other jobs are in flight — yours and everyone else’s. Browser tabs do not speed the farm; closing them also does not pause your job. If slowness is consistent on tiny 15-second 1080p files, check status pages or support, because that is off the typical band. One slow file is a queue. A day of slow files is a capacity event. Either way, do not multiply retries. Watch one job complete before you diagnose the account.

#Queue & Speed
06

Estimating Usage

21.How many credits do I need for a first test?

Budget a structured 6-creative batch — 3 hooks × 2 avatars on one body — in 1080p, plus a small buffer for one or two re-exports of the winner. That is the right first experiment, and the free 50 is meant to be in that neighbourhood rather than a single clip. Add more if you also need a second ratio or a localisation. We will not print an exact credit-per-second table here because it would rot; the confirm step and /pricing are the live numbers. Estimate from length × count × resolution, not from hope.

#Estimating Usage

22.How do I estimate a month of production?

Count finished files, not ideas. If you ship 20 new ads a week at 30 seconds 1080p, plus 25% re-exports, that is the order of magnitude. Add localisation as full extra files. Add 4K only for the masters that need it. Then look at your plan allotment on /pricing and see whether you are an allotment customer or a pack customer. Teams consistently under-count re-exports and over-count “we might test 200 hooks.” Estimate from last month’s ads-manager upload count if you have it. That number is honest.

#Estimating Usage

23.How do I avoid wasting credits on near-duplicates?

Change one variable per batch. Hooks or avatars, not both plus music plus captions plus a new offer. Preview the first two seconds. Kill lines you could paste on a competitor. Cap the first test at six. After you have a winner, generate variants of that winner rather than another dozen randoms. Credits make volume cheap enough that the new failure mode is noise, not scarcity. The discipline is the scarce part. If a file will never be uploaded, do not render it. Editors are free; MP4s are not.

#Estimating Usage

24.Can you tell me exactly how many credits a 27-second ad uses?

Not in this article. The live calculator and the confirm step will. Any number we freeze here will be wrong after the next plan pass, and we will not play that game. What you can rely on: longer costs more than shorter, 4K more than 1080p, each export is paid, failures in our pipeline should not keep the charge, and avatar choice does not change the bill. If a reseller or a screenshot in Slack disagrees with /pricing, /pricing wins. Hedged bands elsewhere on the site are still not a quote.

#Estimating Usage
07

Buying More

25.How do I buy more credits?

Through the in-app plan or pack purchase on /pricing in Klip Kanvas. Buy the plan that matches monthly volume if you render every week; buy packs if you spike. We will not list a dollar figure here. Watch rollover: a giant pack on a plan that only rolls one cycle is how credits expire in a drawer. Agencies should decide whether the house or the client workspace buys. Receipts and seats are billing; this button is capacity. If checkout fails, that is a payment issue, not a render issue — fix the card before you queue twenty jobs.

#Buying More

26.Is it cheaper to upgrade the plan or to buy packs?

If you render most weeks, the plan allotment is usually the cleaner math because you also buy gates — seats, 4K, priority, cloning — not only credits. Packs are for spikes and for staying on a plan you otherwise like. We cannot tell you a break-even dollar amount without violating the no-price rule, and it would be wrong next quarter anyway. Sketch your monthly finished-file count, open /pricing, and pick. Do not buy both in a panic in the same afternoon without checking rollover. Stacked leftover credits still only roll one cycle.

#Buying More

27.Can I get more credits if a campaign is on fire this week?

Yes, by buying a pack or upgrading, not by asking support to “just add some.” Support can help with failed-job refunds and stuck queues; they are not a discretionary credit window for media performance. Plan for spikes if you know a product drop is coming. Rollover exists so a quiet week can fund a loud one, once. If you are consistently surprised, your estimate is wrong — raise the plan. Speed of purchase is minutes; speed of render is still 3–5 minutes a file. Buying more does not make the farm skip the queue.

#Buying More

28.Do promotional or referral credits behave like plan credits?

Not always. Promos can expire faster, exclude 4K, or sit outside the one-cycle rollover. Read the promo terms at grant time. Do not build a production plan on a referral bonus. Spend promo credits on tests, keep plan credits for the winners you must re-export. If a promotion looks too generous on a third-party coupon site, confirm in-app. We will not honour a blog’s invented credit dump. When promo and paid credits coexist, the product should spend in a defined order — check the wallet UI rather than assuming FIFO from memory.

#Buying More#Rollover & Expiry

Ready to put this into practice?

Create your first AI UGC video ad in minutes — no filming, no actors, no editing.

Try Klip Kanvas free

More in this section

16 min read

Veo, Runway and Kling vs AI UGC Ad Tools FAQ

Cinematic generators like Veo, Runway and Kling are not a product-URL-to-UGC-ad factory. Compare jobs, cost per usable ad, URL workflow, and when Klip Kanvas is the ad desk.

FAQAI UGC
Read more
15 min read

Perplexity, Copilot and DeepSeek FAQ for Marketing Teams

How marketing teams should use Perplexity, Microsoft Copilot and DeepSeek for research, office work and cheaper inference — and why none of them replace a UGC ad desk.

FAQAdvanced
Read more

Ready to make ads like these?

Paste a product link and Klip Kanvas writes the script, casts the creator and renders the ad — no filming, no actors, no editing.

Try Klip Kanvas free