OpenArt Review

Sub-Scores by Criterion

• Privacy (30%)7.6 / 10

  No training on uploads, creations private by default, account deletion in settings, DMCA agent with a street address, 18+ terms. Deductions: inputs may be shared with third-party model providers, and non-subscribers’ creations are deleted after 7 days without a per-image control.

• Quality (25%)8.4 / 10

  100+ models, style presets and a LoRA trainer that reproduced our test subject’s likeness in 4 of 5 prompts after a 12-image training run; consistent-character tools and 5–10 second video. Deduction: the NSFW filter stopped 2 of 5 SFW prompts once each.

• Price (20%)7.0 / 10

  Unlimited basic image generations on the plan without payment, but with a possible watermark and 7-day retention. Paid plans from $13 per seat per month; commercial use only from Plus at $27; refunds generally not allowed; add-on credits only with a subscription.

• Speed (15%)8.0 / 10

  5 timed image generations returned in 6–18 seconds, median 11 seconds; the 12-image LoRA training run finished in about 9 minutes. Paid tiers add 8 to 32 parallel generations.

• Support (10%)8.2 / 10

  Help center with dated articles, e-mail support with a stated content-violation channel, Discord community, priority support on Pro. Deduction: no public API and a strict refund stance.

Weighted total: 7.6×0.30 + 8.4×0.25 + 7.0×0.20 + 8.0×0.15 + 8.2×0.10 = 7.8 → 7.8/10. Weights per How we rate: privacy 30%, quality 25%, price 20%, speed 15%, support 10%.

• Session: August 29, 2026, own account with a dedicated editorial email address on the plan without payment, later a 1-month Starter subscription at $13 for the LoRA test; 5 style-transfer prompts and 1 LoRA trained on 12 photos of a consenting editorial team member per our Responsible AI policy. Clothoff AI accepts no uploads — every run happened inside the reviewed tool.
• Speed: 5 timed image generations returned in 6–18 seconds, median 11 seconds; the 12-image LoRA training run finished in about 9 minutes; a 5-second video clip took just under 2 minutes.
• Price: Starter $13 per seat per month with 4,000 credits and 8 parallel generations; Plus $27 (12,000 credits, commercial use); Pro $44 (24,000 credits, priority support); Wonder $175 (106,000 credits); add-on credits $15 per month, subscription required. Prices shown as promotional on September 3, 2026.
• Retention per policy: subscribers’ creations kept until deleted; non-subscribers’ unpublished creations deleted after 7 days; account deletion in settings; no training on uploads (help center); inputs may be shared with model providers (terms). Accessed September 3, 2026.
• Consent tooling: terms require 18+ and prohibit deep fakes and sexually explicit content; an automated filter stopped 2 of our 5 SFW prompts once each with a risk-control message before a successful regeneration. No consent prompt at upload.
• Quality: the LoRA reproduced the subject’s likeness in 4 of 5 prompts; 1 prompt drifted in hair color. Style presets applied cleanly to all 5 control images; 1 of 5 outputs on the plan without payment carried a visible watermark.

Retest SLA: every score is re-tested within 90 days of publication — next check by December 2, 2026. Corrections: editorial@clothoff.ai · Editorial policy.

Prices are in U.S. dollars as displayed on September 3, 2026 and can change without notice; the provider’s own page is the binding source. Our editorial policy requires a re-check within 90 days.

A Good Fit If

• You want to build a consistent character or art style with your own photos and consent.
• You need many models under one subscription with published credit counts.
• You value a stated no-training policy and account-level deletion over anonymity.

Look Elsewhere If

• You need adult or explicit output — OpenArt filters and prohibits it; compare the Pornworks review instead.
• You expect refunds or a long retention window without paying for a subscription.
• You want a photo-edit tool rather than a generator — see the AILabTools review.

Privacy carries 30% of the score. OpenArt earns 7.6 for a help-center statement that uploads are not used to train models, for private-by-default creations, for account deletion in settings and for a DMCA agent with a street address in the terms. Deductions: the terms allow inputs to be shared with third-party model providers to produce outputs, and non-subscribers’ unpublished creations are deleted after 7 days without a per-image retention control — a loss risk rather than a leak risk.

“Non-subscribers’ private creations are automatically deleted after 7 days if they have not been published.”

— OpenArt terms of service, data retention section, accessed via /go/openart-ai/ on Sep 3, 2026

For LoRA training the privacy question is about likeness: 12 photos of one person become a reusable model, so consent has to be explicit and written. Our own LoRA guide explains the workflow and the consent checklist we used, and the art gallery shows SFW style samples from reviewed tools.

This section is informational and not legal advice. OpenArt requires users to be 18+ and prohibits sexually explicit content and deep fakes in its terms, which places it outside the adult-tool category, but U.S. law still applies to the inputs: training a model on another person’s likeness without consent can violate right-of-publicity statutes in states such as California, New York and Texas. Under the TAKE IT DOWN Act (Pub. L. 119-12, signed May 19, 2025), publishing a non-consensual intimate image of a real person is a federal crime, and since May 19, 2026 covered platforms must remove reported images within 48 hours; the FTC can seek civil penalties of $53,088 per violation. A growing list of states also requires age verification on adult sites — see state age verification. Only use photos of yourself or of consenting adults 18+ (Responsible AI); to report an intimate image, use Report NCII.

“a covered platform shall, as soon as possible, but not later than 48 hours after receiving such request— (A) remove the intimate visual depiction; and (B) make reasonable efforts to identify and remove any known identical copies of such depiction.”

— TAKE IT DOWN Act, Pub. L. 119-12, §3(a)(3), GovInfo, accessed September 3, 2026