Claid.ai

By Minel Gunesoglu, founder. I read Claid.ai's terms of use end to end on 13 August 2026 and counted the words in it rather than skimming for them, opened the pricing page in a browser and switched both billing views by hand because a plain fetch of that page returns no price at all, and went through all eighteen of Claid.ai's Trustpilot reviews one at a time. No account was opened, no photograph was uploaded and no picture was produced, so no sentence below describes what Claid.ai makes. Last updated 13 August 2026, re-checked monthly.

Claid.ai's terms of use run to 35,009 characters. The word "output" appears in none of them, and neither does the word "generated". That is a count rather than an impression: the document was fetched on 13 August 2026, saved, and searched (claid.ai/terms). So the agreement a Claid.ai customer accepts at sign-up does not address who owns a picture the tool makes, while the plain-language answer to that question sits on a different page, in a marketing FAQ, where it reads: "No, images generated using Claid are not subject to copyright."

Claid.ai is an image editing and generation suite for online sellers, sold at claid.ai. Claid.ai's own menus name background removal, an enhancer and upscaler, an object eraser, an image expander, AI Photoshoot, AI Fashion and an image-to-video tool, and every one of those starts from a photograph the seller already has. One qualifier belongs beside that: the same page's API menu also lists a Text-to-Image endpoint, so the product is not purely an editor at the developer layer.

The product is published by Lets Enhance Inc., named in the first paragraph of the terms and again in the pricing-page footer, and that company also publishes a separate photo-enhancement product called Let's Enhance. The company's Wikidata entry is filed under the sibling product's name rather than Claid's, which is one reason the two get confused: two third-party posts read while researching this page present Claid.ai and Let's Enhance as alternatives to one another, with no sign either author knew they share a parent.

Claid's Commercial Terms: the Contract Never Says Who Owns a Generated Image

Section 4 of the terms is the ownership clause, and reading it closely is the whole exercise. Section 4 leaves the customer owning everything "collected, downloaded, or otherwise received, directly or indirectly, from you (or on your behalf) by or through our Services", defines that as "Your Content", and then states in capitals: "WE CLAIM NO OWNERSHIP RIGHTS OVER YOUR CONTENT". Every word of that is about material moving from the customer into the service. The photograph a seller uploads stays the seller's. What Claid.ai's tools make out of that photograph is a different thing, and the definition does not reach it.

Nothing else in the document reaches a generated picture either, which is why the word count matters. Searching the saved text for "output" returns zero hits, and searching it for "generated" returns zero hits. The nearest strings anywhere in the agreement are "generation of training data" and "generating invoices", both inside the acceptable-use list, and neither is about who owns a picture. A 35,009-character contract with no sentence on the subject is not an ambiguous sentence to be interpreted; it is an absence.

The pricing page answers the ownership question in one line, and where that line sits is the point. Under "FAQs", verbatim: "Are the images generated or edited subject to any Claid copyright? No, images generated using Claid are not subject to copyright." (claid.ai/pricing, retrieved 13 August 2026.) The two documents are not in conflict, because they are not answering in the same place: one is marketing copy on a sales page and the other is the agreement. This page states both and resolves neither, and the score card records the position as unclear for exactly that reason.

Three further positions in the same agreement are settled rather than silent, and a business reader should price them in. Liability is capped by section 15 at "THE AMOUNT YOU PAID TO US HEREUNDER OR $100.00, WHICHEVER IS LOWER", in a document that also covers a Business tier sold with "Advanced API workflows" and "Custom SLAs and setup". Section 17 requires the customer to "defend, indemnify and hold us and our affiliates, agents, suppliers, or licensors" harmless across seven listed situations, the first of which is simply "your access to or use of our Services", and no clause anywhere sends an equivalent obligation the other way. Section 3 asks users not to use the service "for research and development of artificial intelligence (generation of training data)" or "for benchmarking or competitive analysis of our Services", which a team should know before planning either.

Two clauses address where a dispute would be heard, and they name different arbitration bodies. Section 18 sends disputes to binding arbitration under American Arbitration Association rules, seated in the County of Los Angeles, California, and waives class actions and consolidated proceedings. Section 19, further down the same page, says disputes are "finally settled in accordance with the Comprehensive Arbitration Rules of the Judicial Arbitration and Mediation Service, Inc." by three arbitrators, also in Los Angeles. Both clauses sit in the one agreement, and this page does not judge which of them governs.

Claid's Free Trial and Commercial Use: a Question the Terms Never Answer

Whether work made on the free trial may be sold is the question a seller asks first, and the contract does not take a position on it. Reading the whole document on 13 August 2026 and searching it three separate ways for a tier-specific limit (free or trial near commercial, the bare phrase non-commercial, and personal followed by non), turned up no such clause at all. That is a dated negative and it is not a permission, because a contract that says nothing has not said yes.

The marketing FAQ quoted above does not draw a tier line either. Its answer about copyright is written about images generated using Claid, without distinguishing a trial account from a paid one. A seller who needs certainty before a picture goes on a live listing has one route that does not depend on interpretation, which is to get the answer from Claid.ai in writing and keep it.

Claid's Refund Policy: No Cooling-Off Window, and No Path a Customer Can Start

Refunds come up three times in the terms and every appearance points the same way. Section 4 ends with "Claid will have no obligation to provide a refund of any amounts previously paid". Section 13 says accounts believed to be in violation may be terminated "without notice and without refund". The one clause that does promise money back runs outward from the company: "if Claid terminates your paid Services account prior to its expiration, Claid will refund you the pro rata apportionment of any unused funds", which is a refund a customer cannot invoke.

Cancelling is straightforward, and separate from getting money back. Section 13 lets either party end a paid account "upon thirty days prior written notice for any reason", and adds that a user who wants out "may simply discontinue using the Services". Neither a cooling-off window nor a money-back guarantee appears anywhere in the agreement. Claid.ai's help centre carries its own refund article, which was not captured for this reading, so what this page reports is what the contract says and nothing beyond it.

Claid's Score Card: What Its Four Rows Measure, and Which One Is Not the Axis It Sits On

Four cells carry a figure. The top one changed its name on 19 August 2026, for a reason that has nothing to do with this vendor.

Multiply 84.5, 74.5, 86.3 and 57.9, take the fourth root of the 31,455,917 that comes out, and the answer is 74.9, which is the number this page publishes, 3.7 on a five-point scale, and the middle band of the five rather than the top one. The contract cell is still doing the pulling. Raise it to the level of the usability cell and the same operation returns about 80; hold it where it is and near-perfect readings in the other three would land the composite around 84. On this model a document costs more than a feature does, and a document is the one input a vendor can change in an afternoon.

The top row is the one to read carefully, because its position invites the wrong reading. It sits where output quality sits on cards that have one, and it is not output quality. Delivery envelope asks what an Essentials subscriber can take away (the size, the file types, the route out) and answers nothing about whether the picture is any good. Claid.ai edits a photograph the seller supplies, so a 64-megapixel ceiling caps what the tier will hand back rather than describing detail the tool created; a seller who uploads a small phone photo does not get 64 megapixels of anything.

Two of that row's three readings carry a qualifier that has to travel with them. The resolution figure comes off a per-plan table that exists on Claid's help centre only as a screenshot with an empty alt attribute, which no text gate can read and which is dated 27/08/2024 and labelled "Claid Essential" where the page now sells "Essentials"; the raw image is saved as its own manifest entry beside the transcription, and what makes the figure citable is the live pricing page printing "Up to 64 MP" for Essentials two vendor surfaces and twenty months apart. The API reading is a published Batch API reference with real delivery mechanics, and it earns its value because the API is metered separately rather than shut off, the plan table carries "Claid API URL" and "Claid API Storage" as their own rows and the pricing page sells an API starter pack. Nothing in any of that establishes that an Essentials subscriber gets bundled API access, and no sentence here should be read as saying so; the bundled key sits at Pro.

The fourth reading is missing, and the way it is missing is the point. Claid publishes nothing about watermarking its output, established by exhaustion rather than by sampling: the sitemap names 100 English help articles, all 100 were fetched and searched, and the single match is an article about erasing watermarks from the seller's own photograph. There are no hits across the terms, the pricing page, the API pricing page or either docs page. Silence is not a denial. This page does not claim Essentials output is unwatermarked; an unanswered position leaves the average rather than being scored as a pass.

Two of the remaining cells deserve a caution rather than a defence. Eighteen reviewers is a small base to read a usability figure from, and the section below says what that base does and does not contain. The value cell is entered on the annual-equivalent price because that is the basis the sibling card in this category uses, so the two numbers describe the same kind of purchase; Essentials month to month is $15 rather than $9, and both figures appear in full further down. Not one of the four cells touches what a picture looks like when it comes out. That is the rules a number on this site has to satisfy working as intended rather than a gap in this particular research, and it also means 74.9 is not a mean over the four dimensions this model names, since one of the four is answering a question its own heading disclaims.

Claid Pricing: Essentials and Pro on Both the Monthly and the Annual View

Start with why this section required a browser. A plain fetch of claid.ai/pricing returns one dollar figure in the entire served document, a "Save $120" badge attached to no plan, and neither self-serve tier's price is in those bytes at all. The prices render in the page. So the page was opened, the billing switch was moved by hand, and both states were read and saved. The switch arrives set to yearly, which is the lower of the two numbers a buyer will be quoted.

Every cell above was read from claid.ai/pricing on 13 August 2026, in both switch positions, and the page prints its own warning under the table: "*Prices might be subject to change". Only the Essentials and Pro digits differ between the two views; the rest of the page is identical in both.

Divide each price by its credits and the two paid tiers turn out to cost almost the same per unit on the annual basis, at 1.8 cents a credit for Essentials and 1.75 cents for Pro. On the monthly basis the two separate: 3 cents against 2.45 cents, so a buyer paying month to month gets a real per-credit discount for stepping up, while a buyer paying yearly is buying ceilings rather than a better rate. That arithmetic is this page's, run on the vendor's own four prices and two credit allowances.

The ceilings do not move in the direction a buyer expects. Claid.ai's own feature table caps uploads at 48 MP on the free trial, 24 MP on Essentials and 48 MP on Pro, and caps downloads at 128 MP on the trial, 64 MP on Essentials and 128 MP on Pro. Read across those two rows and the first paid tier sits below the free trial on both limits, so a seller who evaluated at trial resolution and then bought the cheaper plan is working to half the pixel ceiling they were shown. Pro's own plan card meanwhile advertises "Higher upload limits up to 64 MP" while the feature table beneath it gives Pro 48 MP, and nothing on the page ties those two numbers together. Business is listed as "Custom" for uploads and 559 MP for downloads. History retention moves the ordinary way: 24 hours on the trial, one month on Essentials, three months on Pro.

One badge should not be turned into arithmetic. The "Save $120" line sits above the billing switch rather than on either plan card, and the annual saving each tier carries works out differently: $72 a year for Essentials and $168 for Pro, on the four prices the page publishes. Neither figure is $120, so the badge is a headline rather than a calculation about a plan a buyer is choosing between.

Claid's Free Trial Is 50 One-Time Credits, Not a Recurring Free Plan

The free tier is labelled "Free trial" on the card itself rather than a free plan, and what it holds is "50 credits to use on any operation" plus "50 API credits to explore Claid API". The paid tiers are described in the FAQ as a "monthly credit allocation"; nothing in the captured page describes the trial's 50 credits refreshing on any schedule. For a seller arriving from a category where a free tier is a standing monthly allowance, that is a structural difference rather than a smaller number.

The same page also describes the trial a second way. Under a "Try before you buy" panel further down it reads: "Get familiar with Claid and all the features from our Professional plan with 5 trial images". Fifty credits and five trial images are two different units, both live on one page on one day, and nothing on the page ties them together.

How Claid Credits Work: What Background Removal, Upscaling and AI Photoshoot Cost

Claid.ai publishes a per-operation rate card, which makes the cost of a month's work computable before anyone subscribes rather than after. The right-hand column below is Claid.ai's own figure for what the 50-credit trial covers rather than this page's division, and it reconciles against the credit rate in every row.

Two of those rows carry a variable the table does not show. Upscaling is quoted as a range because its cost depends on image size, so the 25 figure is the cheapest end of that range rather than a flat allowance. AI Photoshoot is listed at 4 credits in the feature table, while the FAQ on the same page gives two rates for it: Standard Quality at 4 credits each and Studio Quality at 10. At the higher setting the trial's 50 credits buy five generations instead of twelve, which is arithmetic on the vendor's own pair of rates, and the only thing the buyer changed was a quality setting.

One Claid AI Video Costs 35 of the Trial's 50 Credits

Video is metered on a completely different scale from everything else on the rate card. One five-second AI Video costs 35 credits, and Claid.ai's own feature table puts the trial's allowance for that operation at exactly one. So a trial account gets a single clip and keeps 15 credits, which is fifteen background removals or three AI Photoshoots at Standard Quality. On the paid tiers the same operation yields 14 videos a month on Essentials and 57 on Pro, again the vendor's own figures.

Whether that video is produced by Claid.ai's own model or by a bundled third-party model is not stated anywhere on the captured page, and this page does not guess. Nothing here evaluates the clip either, for the reason the score card gives: no picture was produced for this review.

Claid for Shopify and Catalogue Integration: an API and Zapier Path, and No App-Store Listing Found

Claid.ai's integration story is developer-shaped, and that is the single fact most likely to change a shortlist. The vendor's own footer routes integration through a "Claid.ai x Zapier" page, an "Enterprise API" page and a "Custom workflows" page, and its industry pages address online marketplaces, ecommerce, print on demand, food, fashion, automotive and real estate rather than named consumer storefronts. The pricing table's export row names "Instagram stories, Shopify products, and more" as destinations for a finished file.

Does Claid Have a Shopify App? the Integration Path Is REST API or Zapier

No Shopify App Store listing for Claid.ai surfaced in this desk's searching on 13 August 2026, and the routes the vendor documents itself are a REST API and a Zapier connection. For a shop owner who expected to install something and click through a setup screen, that is a different purchase: an API or automation route needs somebody who can wire it up, whether that is an in-house developer, an agency or an existing automation-platform account.

The consequence is worth stating concretely. A seller with a few dozen listings and no technical help can still use Claid.ai through its web app, uploading and downloading by hand, and the credit rates above tell them what a month of that costs. What they will not get from that route is the thing the API is sold for, which is images moving between a store and the tool without a person in the middle.

Claid's API Pricing: $59 for a Thousand Credits, on a Meter That Does Not Match the Web App

Claid.ai sells API access on its own page, claid.ai/api-pricing, read on 13 August 2026, and it is not a subscription. A thousand credits cost $59 as a one-time purchase, about six cents a credit, which is arithmetic on the vendor's two numbers rather than a rate the page prints. Requests are capped at 4 per second and 120 per minute by default, and the page says some generation endpoints sit lower than that.

The part worth knowing before a team sizes anything is that the two meters are not the same meter. Claid.ai's own FAQ puts it plainly: "They're two separate credit types with separate pools and pricing. Web (subscription) plan credits are used in the Claid web app; API credits power the image and video API and are billed per operation. They aren't interchangeable, and one can't be spent on the other." The same operation is priced differently on each side. Removing a background costs one credit in the web app and two through the API. The web app's AI Photoshoot row costs four credits; the API's AI backgrounds row, its nearest equivalent, costs three. Upscaling runs 2 to 8 credits in the web app and 1 to 11 through the API, because the API bills on the size of the picture that comes out. Only video matches: 35 credits for five seconds either way, and 70 for ten on the API.

So a developer who reads the web app's rate card and multiplies it by an expected monthly volume is working from the wrong table, and the direction of the error depends on which operation the pipeline leans on. The trial does reach both surfaces (the web pricing page grants "50 API credits to explore Claid API" alongside the web credits) but that is the only place the two systems touch.

One row on the API list is worth flagging for a different reason. The operation named Text-to-image, at one credit, creates a picture from a written prompt with no photograph involved at all. Everything else this page describes starts from a picture a seller already owns, and that distinction is the one this site uses to decide what belongs in this category. The row is stated here as a fact about the vendor's list, not resolved.

Claid Reviews: an Unprompted Trustpilot Base, and No Reddit Thread in Four Years

The public record on Claid.ai is small, lopsided and concentrated on one platform, and the honest way to use it is to say what it samples. Trustpilot carried enough reviews to score on 13 August 2026; the other platforms read that day did not, for reasons worth listing once. G2 refused every retrieval method attempted, so no figure from it appears here. Capterra's correct listing returned 404 and 403 on repeated attempts, while a different product of the same name renders fine, so no Capterra figure appears either. Product Hunt carries a listing with reviews, but that platform prints relative ages such as "4 years ago" instead of dates, so nothing from it is quoted or counted below. Hacker News holds two entries, both 2022 announcement submissions from the company itself, with no organic discussion attached.

Claid's Trustpilot Sample: 18 Unprompted Reviews, Not a Solicited Base

Trustpilot rates Claid.ai 4.1 out of 5 across 18 reviews, and the distribution is the unusual part: 94% five-star, 6% one-star, and 0% in each of the four-, three- and two-star bands (trustpilot.com/review/claid.ai, read 13 August 2026). A shape like that invites a conclusion about how the reviews were gathered, and the page rules that conclusion out in its own words. Printed on the profile: "No history of asking for reviews" and "This company hasn't invited their customers, so reviews may not be representative". Every one of the 18 also carries Trustpilot's "Unprompted review" label. Self-selected and vendor-solicited are different claims, and only the first is supported here.

Two facts about timing keep 18 from being read as eighteen steady signals. Trustpilot's own counter says six reviews in the last twelve months, and six of the 18 were posted in the first three weeks of June 2024. The profile is also unclaimed, which means there is no company reply beside any review on it, so nothing quoted below has been answered.

Counting the reviews one at a time turned up the finding the aggregate hides: nine of the 18 praise customer support or a named human on live chat rather than the pictures. Half a review base for an image tool is describing the people behind it.

"What really helped a lot was the online chat, because it was a conversation with a human being (I love it!) familiar with the tool and able to understand my use case." (Ursula Weinmann, five stars, Trustpilot, posted 18 February 2026)

"I was struggling with a storage config at first and reached out to customer service through their chat widget. They got back to me very quickly and were able to guide through solving every obstacle in my way." (João Guilherme S. Morossini, five stars, Trustpilot, posted 6 June 2024)

A second cluster describes the API in a developer's own vocabulary rather than a marketer's. "I've picked this solutioin due to API functionality. API is really easy to use, has a lot of options and settings," wrote Anton H. on 19 June 2024, spelling as posted. "The API is invaluable for image quality automation," wrote Art Altmäe on 11 July 2024. A third and separate group is not technical at all: "I used to spend hours gathering images and formatting product shots to create benefit 'cards' for my online listings. Not anymore!" wrote Jen O'Farrell on 3 February 2025. Two audiences, one review page, and the tool is landing with both.

On the question a buyer most wants settled, the same profile answers in two opposite directions about six weeks apart. All dates here are Trustpilot's posting dates; the platform prints a separate date of experience under each review, and the two are not the same field.

"Surprised by all the positive reviews. The tool's outcomes always look ridiculous and super unnatural." (O Whitcombe, one star, Trustpilot, posted 10 May 2026)

"you blink once and you cannot tell the difference between what's real and what's not" (Kongregate Kong, five stars, Trustpilot, updated 27 March 2026)

Nothing on this page adjudicates that. No picture was produced for this review, no arena measures the category, and two named reviewers on one platform reached opposite conclusions about the same thing. Reporting the disagreement is the whole of what the evidence supports.

Claid on Reddit: No Thread Found Across Six Search Angles and Four Years

Reddit returned nothing about Claid.ai across six independent search angles on 13 August 2026: a general review search, a billing search, an API-experience search, a background-removal search, subreddit-scoped searches inside seller communities, and a bare site-restricted sweep, covering the product's whole public history since its 2022 launch. Anyone running that search themselves should qualify it with the domain, because an unqualified query for Claid on Reddit returned six of ten results about Claude, Anthropic's assistant, on the same day.

An absence like that is worth stating rather than skipping, and it is also worth not over-reading. What it means is that the sentiment which does exist for Claid.ai sits on platforms where a user arrives already intending to review something, and that no independent thread exists to check the Trustpilot picture against.

Claid Alternatives: the Names Sellers Reach For, and Why No Score Sits Beside Them Here

The trap in this particular shortlist is not a rival at all, and a seller researching Claid.ai will hit it within about two searches. Let's Enhance, at letsenhance.io, is a different product published by the same company, Lets Enhance Inc., and it carries its own separate review profiles on the same platforms. Reviews written about Let's Enhance are not evidence about Claid.ai, and none of them is borrowed anywhere on this page.

Photoroom is the peer that carries a dated card on this site, and one documented difference matters more than any ranking would: Photoroom's own dated card records a free plan licensed, in that vendor's published words, for personal and non-commercial purposes, which is precisely the question Claid.ai's contract leaves unanswered in either direction. Pebblely, at pebblely.com, turns up in the same third-party roundups that name Claid.ai and carries no card here.

Where a tool has been scored, its number lives on its own card and on the category page that holds this site's product-photography cards. No score on this page is set against another score, and none of the names above is placed above or below Claid.ai here.

Is Claid Worth It? What an Average Composite Means for a Seller's Decision

The buyer this fits can be described precisely, which is more useful than a verdict. Someone whose route in is the API or an automation connector rather than an app install; who will size a month of credits against the per-operation rates before subscribing rather than after; who wants pixel ceilings above Essentials and is therefore looking at Pro from the start; and who can live with an agreement that does not answer the ownership question, or who intends to get that answer from Claid.ai in writing before anything goes on a listing. For that buyer the support record is a genuine argument in the tool's favour, and it is what Claid.ai's public record repeats most often.

Hold off in four situations, and the first of them is probably why you are reading. If you need the picture-ownership answer to be in the contract, you will not find it there, in a document effective 10 August 2023 that carries no amendment date. Anyone budgeting from the first price they see is reading the yearly rate, because that is where the switch starts; the month-to-month figures are $15 and $49. Anyone planning to evaluate on the trial and then buy Essentials should check the upload and download ceilings first, since both drop below what the trial allowed. And anyone whose purchase depends on being able to get money back should read section 13 before paying, because the only refund the document promises is one the company initiates.

The one question 74.9 is silent on happens to be the first one a seller asks: does the finished picture look good enough to put on a listing. All four rows carry a number and none of them answers that, because nothing published anywhere measures it. Of the four, the contract reading is both the lowest and the most fixable, and it is the one that decides the composite. Everything else on this page can be re-checked on three of Claid.ai's own pages in about ten minutes: a pricing page that keeps its prices out of its own source, a rate card that makes a month computable, an API rate card metered separately from that one, and an agreement whose most important sentence about generated images is the one that is not in it.

Written by Minel Gunesoglu, founder of Vouch. Sources for this page were saved as bytes on the day they were read: Claid.ai's terms of use, its pricing page in both billing states, its API pricing page, and its Trustpilot profile, all on 13 August 2026. Every figure above carries that date where it appears, and prices in this category go out of date faster than anything else on a review page. More on who writes these pages, and on what evidence · LinkedIn. Disclosure: Claid.ai runs an affiliate programme and this site is not in it, so every vendor link above is a plain link; affiliate links elsewhere on this site are marked, and a paid submission buys review speed only, never a listing and never a score.