Does your AI ad carry a watermark? Tool by tool
An AI watermark is anything a tool leaves in your file to show that AI made it. It comes in three forms: a logo you can see in the corner of the frame, a label hidden in the file’s data (C2PA Content Credentials, a sealed record that travels with the file), and an invisible signal woven into the pixels or the sound itself, such as Google’s SynthID.
We read the official pages of 63 AI ad tools on 12 October 2026. On 21 of them, free or trial exports carry a visible logo. 9 add the hidden label to at least some files, 5 say their files carry SynthID, and 19 say nothing about any of the three.
Paying usually gets rid of the logo: 16 of those 21 tools drop it on a paid plan. Paying does not remove the invisible signal. Kling, Higgsfield and Google say theirs stays, and no tool says it goes. That matters when you use Veo or Nano Banana inside another app: 34 tools resell them, and 2 of those say Google’s SynthID still ends up in your file.
8 of the 12 video generators mark free exports; Grok Imagine keeps its mark on every plan.
| Tool | Visible mark | Removed on | C2PA | Invisible mark | Source |
|---|---|---|---|---|---|
| Google Veo | A setting in Flow | Setting off, except in India, South Korea and Vietnam | Yes | SynthID | Flow Help, undated |
| Kling | Free and non-member credits | Any membership | Unclear | Unnamed, every output | Paid terms 7.2, 10 Oct 2026 |
| Runway | Free | Standard and up | Yes | Not stated | Help centre, 30 Jul 2026 |
| Luma | Free and Lite, for good | Paid plans, new work only | Yes | Not stated | Licensing guide, undated |
| Hailuo | Free | Standard and up; credit packs do not | Not stated | Not stated | Subscription terms, 14 Jul 2025 |
| Seedance (Dreamina) | Free | Unverified | Yes | Unnamed (ByteDance) | Seedance 2.5 post, 31 Jul 2026 |
| LTX Studio | Free | Lite and up | Not stated | Not stated | Pricing, undated |
| Grok Imagine | All plans | No setting removes it | Not stated | Not stated | xAI FAQ, undated |
| Vidu | Unclear | Not stated by plan | Not stated | Not stated | Terms 5, 3 Jul 2026 |
| PixVerse | Unverified | Unverified | Not stated | Not stated | Plans page, undated |
| Topview | No (no free plan) | Every plan | Not stated | Not stated | Pricing, undated |
| Pippit | Free | Starter and up | Not stated | Not stated | Pricing, undated |
Scroll sideways for all 6 columns
6 of the 10 AI UGC and avatar tools mark free or trial exports; D-ID marks its Pro plan too, and TikTok Symphony labels every export.
| Tool | Visible mark | Removed on | C2PA | Invisible mark | Source |
|---|---|---|---|---|---|
| Creatify | Free | Standard and up, per video | Not stated | Not stated | Help centre, 19 Aug 2026 |
| HeyGen | Free; on by default on paid plans | Creator and up, a toggle | Not stated | Not stated | Help centre, undated |
| Synthesia | Basic (free) | Starter and up | Yes | Not stated | Help centre, 13 Aug 2026 |
| MakeUGC | No (no free plan) | Every plan | Not stated | Not stated | FAQ, undated |
| JoggAI | Unverified | Starter and up | Not stated | Not stated | Pricing, undated |
| Captions | Basic, with AI media | Max and up | Not stated | Not stated | Help centre, 21 Sep 2026 |
| D-ID | Trial, Lite and Pro | Not stated | Not stated | Synthetic mark, every plan | EULA 4.2, Jul 2024 |
| TikTok Symphony | AI label on every export | Not removable | Not stated | Not stated | Ads help, Jun 2026 |
| Tagshop | No (no free plan) | Every plan | Not stated | Not stated | Pricing, undated |
| Nextify | Free | Starter and up | Not stated | Not stated | FAQ, undated |
Scroll sideways for all 6 columns
5 of the 11 creative suites mark free exports, and each of the 5 drops the mark on a paid plan.
| Tool | Visible mark | Removed on | C2PA | Invisible mark | Source |
|---|---|---|---|---|---|
| Zeely | No | Every plan | Not stated | Not stated | Zeely blog, 10 Oct 2026 |
| PromptWise | Not stated | Not stated | Unclear | Not stated | Trust and Safety, 28 Sep 2026 |
| Higgsfield | Free | Any paid plan, new work only | Yes on Gemini Omni Flash | Unnamed markers; SynthID on Omni Flash | Help centre, 3 Aug 2026 |
| Krea | No on Free | Not stated for paid plans | Not stated | Not stated | Krea blog, 26 Aug 2026 |
| invideo | Free | Starter and up, after a re-export | Not stated | Not stated | Help centre, 23 Sep 2025 |
| OpenArt | Free | Starter and up | Not stated | Not stated | Terms 4.5, 30 Jul 2026 |
| Pika | On links shared from Pika | Downloads, every plan | Not stated | Not stated | FAQ, undated |
| Adobe Firefly | Not stated | Not stated | Yes | Not stated | Adobe release, 24 Apr 2025 |
| Leonardo.Ai | Free, video | Paid plans | Not stated | Not stated | Help centre, 18 Feb 2026 |
| ElevenLabs | Free, video dubs | Paid plans | Not stated | Audio watermark, own models | Pricing, undated |
| Artlist | Unverified | Unverified | Not stated | Not stated | Terms, archived 14 Aug 2026 |
Scroll sideways for all 6 columns
2 of the 11 image tools mark free exports; Pencil marks Nano Banana images on every plan, and ChatGPT and Nano Banana files carry both C2PA and SynthID.
| Tool | Visible mark | Removed on | C2PA | Invisible mark | Source |
|---|---|---|---|---|---|
| Nano Banana | Gemini app, setting on | Setting off; in India, South Korea and Vietnam only on AI Ultra | Yes | SynthID | Gemini Apps Help, undated |
| ChatGPT | Not stated | Not stated | Yes | SynthID | OpenAI provenance guide, undated |
| Midjourney | Unverified | Unverified | Not stated | Unverified | Verify page, undated |
| Canva | Unverified | Unverified | Unverified | Not stated | Help centre, undated |
| Meta Advantage+ creative | Unverified | Unverified | Unverified | Not stated | AI terms, archived Nov 2024 |
| Picsart | Unverified | Unverified | Not stated | Not stated | Pricing, undated |
| AdCreative.ai | No on downloads | Every plan | Not stated | Not stated | Plan table, undated |
| Pencil | Nano Banana images | Not stated | Not stated | SynthID | Pencil blog, 23 Sep 2025 |
| Blaze | Free | Starter and Growth | Not stated | Not stated | Help centre, 20 Mar 2026 |
| Photoroom | No, since Aug 2024 | Every plan | Not stated | Not stated | Photoroom post, 30 Aug 2024 |
| Pixelcut | No on Free exports | Not stated | Not stated | Not stated | Pricing, undated |
Scroll sideways for all 6 columns
| Tool | Pages read |
|---|---|
| Wan | Terms III.8, 6 Aug 2026 |
| Amazon Ads Video Generator | Product terms, 11 Sep 2026 |
| Arcads | Terms, 22 Sep 2026 |
| Hedra | Pricing FAQ, undated |
| Argil | Terms 9.3, undated |
| CreateUGC | Pricing, undated |
| Maxfusion | Terms 15, 19 Jul 2026 |
| Wireflow | Terms 5, 18 Feb 2026 |
| Icon | Terms, undated |
| Magnific | Terms, undated |
| LumiYing | Terms, undated |
| FLORA | Terms, 17 Nov 2025 |
| Ideogram | Terms 6.2, 14 Aug 2024 |
| The Brief | Pricing, undated |
| Omneky | Terms, 27 Aug 2025 |
| Predis.ai | Pricing, undated |
| Claid.ai | Pricing, undated |
| Flair.ai | Pricing, undated |
| Pebblely | Pricing, undated |
Scroll sideways for all 2 columns
What are the three kinds of AI watermark?
The first is the logo you can see: a small brand mark in a corner of the video or image, like Runway’s on its free plan. When a pricing page promises “no watermark”, this is the mark it means, and it is the only one a paid plan removes.
The second is a label inside the file. Content Credentials, built on the C2PA standard, are a signed note stored in the file’s data that names the tool and says AI made it. Think of it as a receipt stapled to the file. Meta and TikTok read it to label posts. It is also easy to lose: OpenAI’s provenance guide warns that “Editing, converting, or sharing a file can remove its metadata.”
The third is a signal you cannot see. An invisible watermark is woven into the picture or sound itself, so it stays when the file is renamed or its data is wiped. Google’s SynthID is the best known, and Gemini puts it, along with Content Credentials, in every image and video it makes.
Turn visible watermarks on or off
1. On your computer, go to gemini.google.com. 2. At the bottom left, click Settings > Media Watermark. 3. Turn the setting on or off.
Tips: This setting only controls the visible watermark. It doesn’t affect SynthID watermarks or Content Credentials for media you create with Gemini Apps; these settings apply only to Gemini Apps (web and mobile), not AI features in other Google products.
SynthID: All AI-generated media created or edited with Gemini Apps includes an invisible SynthID watermark.
Content Credentials (C2PA) provenance signals: Gemini Apps include Content Credentials (C2PA) metadata to provide information about the content’s origin.
OpenAI says SynthID “may survive some transformations”. Kling, ByteDance and ElevenLabs hide marks of their own, and each one can only be read by its maker’s checker: Kling’s detection page checks Kling files, the ElevenLabs Audio Detector checks ElevenLabs audio and Midjourney’s verify page checks Midjourney images.
Does paying remove the watermark?
Yes for the visible logo on 16 of the 21 tools that mark free exports. No tool says paying removes an invisible mark.
The 16 are Kling, Runway, Luma, Hailuo, LTX Studio, Pippit, Creatify, HeyGen, Synthesia, Nextify, Higgsfield, invideo, OpenArt, Leonardo.Ai, ElevenLabs and Blaze. Synthesia’s logo shows on every Basic video until you upgrade and re-generate.
The Synthesia watermark appears on videos generated under certain conditions, such as being on the Basic plan or generating a video in API test mode. This article explains how to remove it in both cases.
If you are on the Basic Plan: The Synthesia logo (watermark) will appear on all videos created with your account. To remove it, upgrade your plan and re-generate the video.
The other 5 do not say paying removes it. Grok Imagine says “There is no setting to remove the watermark”. TikTok Symphony labels every export, D-ID’s Pro plan shows “An AI watermark”, Nano Banana’s mark follows a setting on every plan, and Seedance’s paid removal is unverified. Plan prices sit in the AI ad tool price tracker.
Kling members may remove “brand watermarks” (7.2), yet its detection page says “An imperceptible watermark and metadata are embedded in images and videos”. Higgsfield’s markers “aren’t removed by upgrading” (help centre).
Why does my generation have a watermark?
Paid plans don’t add a visible Higgsfield watermark to your generations, whether they’re credit-based, free generations included in your plan, or Unlimited models. Separately, Higgsfield may embed invisible, machine-readable markers or metadata that identify content as AI-generated: they aren’t a visible watermark and don’t appear on your image or video.
The watermark depends on your account type, not on the model, the credits used, or the output quality. On a free account, every generation carries a Higgsfield watermark. On any paid plan, generations come without a watermark. There is no separate watermark setting to switch on or off. The visible watermark is separate from the invisible, machine-readable markers Higgsfield may embed in AI-generated content; those markers aren’t visible in your file and aren’t removed by upgrading.
Google’s Gemini setting “only controls the visible watermark. It doesn’t affect SynthID watermarks or Content Credentials” (Gemini Apps Help).
Luma’s free-plan mark “cannot be removed, even if you later upgrade”.
Who owns the rights to my generations?
The ownership and usage rights of your generated content depend on your subscription tier:
Free and Lite Plans: Content generated under these plans is for personal use only. You do not have commercial usage rights. Generated content includes a watermark that cannot be removed, even if you later upgrade your subscription.
Plus, Unlimited and Enterprise Plans: Content generated under these plans includes commercial usage rights with no watermark.
HeyGen’s paid plans keep the watermark on until you switch it off, and invideo projects keep theirs until re-exported on a paid plan.
Removing the Watermark as a Paid User
Note: Removing the watermark is available on all paid plans — Creator, Pro, Business, Enterprise, and grandfathered legacy plans. Free user videos automatically include the HeyGen watermark without the option for removal.
If you’re a paid user, you have the option to remove the watermark from your studio videos. Simply click on Generate and the Submit Video menu will appear. Toggle off the watermark option at the bottom left of the menu. Once this is toggled off for one video, it will remain toggled off for all future videos until you turn it back on.
Two resellers say the model maker’s mark reaches your file; the other 32 of the 34 that resell Veo or Nano Banana do not say either way. Google’s Gemini API docs say “All generated images include a SynthID watermark”, and DeepMind’s Veo page says Veo videos carry SynthID.
Higgsfield says so on its Gemini Omni Flash page (“Every output carries SynthID”), and Pencil says it of Nano Banana images (“Every image includes both a visible watermark and invisible SynthID digital watermark”). For Seedance, ByteDance’s API terms bar removing its invisible watermark (BytePlus 2.2).
Removing a mark yourself breaks the terms at Luma (4.9), Hailuo (Terms of Use), Synthesia (Acceptable Use Policy), Argil (9.3), Ideogram (6.2), OpenArt (4.5), Grok (FAQ), D-ID (EULA 4.2) and Wan (II.1(a)). What each plan allows in paid ads is in the rights matrix.
Do Kling, Veo and Higgsfield add a watermark?
Yes, each in a different layer. Kling puts a brand logo on free and non-member work, and an invisible mark with metadata in every output.
Google Veo puts SynthID and Content Credentials in every Flow video. Its visible mark follows a setting, applied automatically in India, South Korea and Vietnam.
Higgsfield marks free work only with a visible logo, and may embed machine-readable markers on any plan. Its Gemini Omni Flash output carries SynthID.
What Meta, TikTok, YouTube and Google Ads read and label
Among ad policies, only Meta’s says it reads C2PA in files from other tools, and it labels the ad AI info when it finds it. No platform page names SynthID among the signals it reads. So a Meta ad cut from a Synthesia, Runway or Google file can carry AI info with no disclosure from you.
| Platform | Signals read | Label shown | Your disclosure duty | Source |
|---|---|---|---|---|
| Meta ads | C2PA; Meta’s own AI tools | “AI info” in About this ad | Political and social-issue ads with “photorealistic” AI media | Business Help, undated |
| Meta posts | C2PA, IPTC, self-disclosure | “AI info” | “Photorealistic” AI video, “realistic-sounding” audio | Newsroom, 6 Feb 2024 |
| TikTok ads | Not stated | AIGC label or your own disclaimer | Fully or largely AI-made ads; undisclosed ones are rejected or restricted | Ads policy, Apr 2026 |
| TikTok posts | C2PA, TikTok’s invisible watermark, detection | AIGC label | “Realistic” AI content | Newsroom, 9 May 2024 (C2PA); Newsroom, 19 Nov 2025 (watermark) |
| YouTube | C2PA, YouTube’s own AI tools, detection | AI label on the player or in the description | “Realistic” altered or synthetic content, at upload | YouTube Help, undated |
| Google Ads | SynthID and C2PA in its own tools’ output; unnamed signals from other platforms | How this ad was made; overlays for the EU, India and New York | Election ads; an optional AI label setting | Google Ads Help, undated |
| C2PA | C2PA icon, rolling out | Optional, in select countries | LinkedIn Help, Oct 2026 |
Scroll sideways for all 5 columns
In some geographies AI labels will also show on the ad itself. For campaigns that target the European Union, India, and New York, ads with assets that you designate as AI created or edited will also include visible overlays on the ads.
Each rule in full: the Meta AI disclosure guide, the TikTok AI label guide and the YouTube AI disclosure guide.
What the law requires, and from when
The law splits the work: the tool maker marks the file, and the advertiser discloses in the ad. Under the EU AI Act, providers “shall ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated” (Article 50(2)).
| Rule | Who it binds | From | What it requires |
|---|---|---|---|
| EU AI Act 50(2) | Tool makers | 2 Aug 2026 | Machine-readable marks on output; tools on the market before 2 Aug 2026 have until 2 Dec 2026 (2026/1744) |
| EU AI Act 50(4) | Deployers | 2 Aug 2026 | Disclose a deep fake at first exposure |
| California AI Transparency Act | Tool makers | 2 Aug 2026 | A latent disclosure naming provider, system, time and a unique ID |
| California AB 2713 | Platforms over 2 million monthly users | 1 Jan 2027 | Show provenance data; never “knowingly strip” it |
| New York GBL 396-b | Advertisers | 9 Jun 2026 | Disclose a synthetic performer in the ad |
| California SB 1050 | Advertisers | 1 Jan 2027 (default start date) | A “clear and conspicuous” synthetic performer disclosure |
Scroll sideways for all 4 columns
SB 1000 (approved 30 September 2026) drops the one-million-monthly-user threshold, so the California Act covers every publicly accessible generator in the state. Under the Act, a tool that only crops or converts a file makes a “minor modification” and owes no latent disclosure for it.
The two synthetic performer laws ask for a disclosure in the ad itself, which a mark inside the file does not give. The AI ad disclosure rules set them beside each platform’s rule.
Method
Each row comes from the tool’s own official pages, read on 12 October 2026. The CSV quotes the line, its clause and its date. C2PA “Yes” also counts a generator listed on the C2PA Conforming Products List.
A tool counts among the 21 when its free or trial plan’s exports carry a visible mark and the page was re-read that day. It counts among the 16 when a paid plan’s page says the mark goes. Captions (free plan unclear), JoggAI and Canva (not re-read) and Pencil (no free plan) sit outside the 21.
“Not stated” means the vendor’s pages say nothing, never that the file carries no mark. “Unclear” means the text allows two readings. “Unverified” means the official page could not be re-read that day.
A correction keeps the row’s anchor and shows in the changelog and on updates and corrections. For terms that bar published comparisons, see benchmark bans by tool.
How to cite this data
To cite one tool, link its row: /data/watermarks/#kling opens the table at Kling. The CSV holds every plan row with the vendor’s quote and clause, the link and both dates.
adtestbench, "Does your AI ad carry a watermark? Tool by tool", https://adtestbench.com/data/watermarks/, updated .
- CSV file /data/watermarks.csv
Sources
- Google: Get started with Google Flow: Google Flow Help, FAQ:
Will my generations contain a watermark?
; undated. Checked - Google: Content Credentials: Gemini Enterprise Agent Platform: Google Cloud Documentation, Supported models (list includes veo-2.0, veo-3.0 and veo-3.1 model IDs); dated 10 October 2026. Checked
- Kling AI: Kling AI Terms of Paid Service, Terms of Paid Service 7.2 (also Terms of Service 4.5); dated 10 October 2026. Checked
- Runway: Free plan details: Runway Help Center, Free plan details; dated 30 July 2026. Checked
- Runway: Runway pricing, Pricing page, Standard card (paid tiers build on Standard); undated. Checked
- Luma AI: Dream Machine Guide: Licensing: Luma AI, Who owns the rights to my generations?; undated. Checked
- Luma AI: FAQ: Luma APIs, FAQ: Watermarks; undated. Checked
- Luma AI: Luma AI Terms of Service, Terms 4.9: customers may not
remove, alter, obscure, or circumvent any such watermarks, metadata, or content credentials
; dated 14 May 2026. Checked - Hailuo AI: Hailuo AI Terms of Use, Bars removing
any watermark, metadata, label, or other notice
identifying AI content; dated 19 August 2026. Checked - Synthesia: Synthesia Acceptable Use Policy, Bars
Removing, deactivating or disabling any watermarks
designed to help validate provenance; undated. Checked - Hailuo AI: Hailuo AI Subscription Service Terms, Subscription Service Terms: Download and Watermark Conditions; dated 14 July 2025. Checked
- Dreamina (CapCut): Giving creators more control with Dreamina Seedance 2.5 and Dola Seedream 5.0 Pro, Pairing powerful creation with responsible innovation; dated 31 July 2026. Checked
- Dreamina (CapCut): Dreamina pricing, Pricing page plan cards; undated. Checked
- Alibaba: Wan Terms of Service, Terms of Service III.8(g)(iii) Your Use of Outputs (II.1(a) bars removing labels or watermarks without permission); dated 6 August 2026. Checked
- Lightricks: LTX Studio pricing, Pricing page, Lite card (the Free card lists no such line); undated. Checked
- xAI: FAQ: Grok Website / Apps (xAI docs), FAQ: Grok Website / Apps: Image & Video Generation (Grok Imagine),
Why do my generated images/videos have a "grok" watermark? Can I remove it?
; undated. Checked - Vidu: Vidu Terms of Service, Terms 5. Prohibited Conduct and Content; dated 3 July 2026. Checked
- PixVerse: PixVerse subscribe, Subscribe page, Basic card (no watermark-free line); undated. Checked
- Topview: TopView pricing, Pricing page,
Compare all features and plans
, Processing & Features row (ticked for all four plans); undated. Checked - Pippit: Pippit pricing, Pricing page, Starter card (plan data lists remove-watermark for Starter, Plus, Pro only); undated. Checked
- Amazon Ads: Amazon Ads Product Terms, Other Terms: Advertising generative artificial intelligence tools (c); dated 11 September 2026. Checked
- Arcads: Arcads Terms, 13.3.1; dated 22 September 2026. Checked
- Creatify: Free Plan, Help centre, FAQs > Free Plan; dated 19 August 2026. Checked
- Creatify: Creatify Pricing, Pricing page, Starter card (Standard card from 16 October 2026); undated. Checked
- HeyGen: How to Remove the HeyGen Watermark, Help centre, How to Remove the HeyGen Watermark; undated. Checked
- Synthesia: How do I remove the Synthesia watermark?, Help centre, How do I remove the Synthesia watermark?; dated 22 July 2026. Checked
- Synthesia: What does the EU AI Act Article 50 mean for my Synthesia videos?, Help centre, What Synthesia Does Automatically > Provenance signals (C2PA); dated 13 August 2026. Checked
- MakeUGC: MakeUGC homepage, Homepage FAQ,
Who owns the videos I create?
; undated. Checked - JoggAI: JoggAI Pricing, Pricing page, comparison table: Free
-
, Starter and up ticked; undated. Checked - Captions: Watermark on Exported Video, Help centre, Troubleshooting > Watermark on Exported Video, step 3; dated 21 September 2026. Checked
- Hedra: Hedra Pricing: Free Plan, No Credit Card, Pricing page, hero line; undated. Checked
- Argil: Argil Terms, Terms 9.3 (User prohibitions); undated. Checked
- D-ID: D-ID Studio Pricing, Pricing page FAQ,
Why do videos have a D-ID watermark on the Trial and Lite plans?
; undated. Checked - D-ID: D-ID Studio End User License Agreement, EULA 4.2; dated July 2024. Checked
- TikTok: How to generate videos with Symphony Creative Studio, Ads help centre, How to generate videos with Symphony Creative Studio, note after step 7
Export your final video
; dated June 2026. Checked - Tagshop AI: Tagshop Pricing, Pricing page, plan cards and comparison table; undated. Checked
- CreateUGC: CreateUGC Pricing, no watermark or provenance statement found; undated. Checked
- Nextify.ai: Nextify.ai homepage, Homepage FAQ,
Do generated videos have watermarks?
; undated. Checked - Nextify.ai: Nextify Pricing, Pricing page, Starter card; undated. Checked
- Maxfusion: Maxfusion Terms, Terms 15; dated 19 July 2026. Checked
- Wireflow: Wireflow Pricing: Free, Starter, Professional & Team Plans, Pricing page, Free plan; undated. Checked
- Wireflow: Wireflow Terms, Terms 5; dated 18 February 2026. Checked
- Icon: Icon Terms, no watermark or provenance statement found; undated. Checked
- Zeely: Is Zeely legit? Honest review for small businesses, section
What Zeely actually is
/ export list; dated 10 October 2026. Checked - PromptWise: Trust & Safety: PromptWise, Trust & Safety s.6
AI-Generated Content and Provenance
; Terms of Service 4.6, 18.1; dated 28 September 2026. Checked - Higgsfield: Why is there a watermark, and how do I remove it?, Help centre article, section
Why does my generation have a watermark?
; dated 3 August 2026. Checked - Higgsfield: Gemini Omni Flash: AI Video Generation (Higgsfield model page), FAQ
What safety and provenance signals are included?
; undated. Checked - Magnific: Magnific Terms of Use, Terms of Use, Magnific Originals Content restrictions; undated. Checked
- Krea: The 7 Best Free AI Image Generators in 2026: Krea, Table
Rights, watermarks, and the fine print
(row Krea); dated 26 August 2026. Checked - Krea: Krea Terms of Use, no watermark or provenance statement found; dated 20 May 2024. Checked
- invideo: Why are there still watermarks on my video if I’m on a paid plan?, Help centre article,
Watermarks on previously generated videos
; dated 23 September 2025. Checked - invideo: invideo pricing, Pricing page plan data (Starter and Plus: no_watermark + full HD; Max and up: no_watermark + 4K); undated. Checked
- LumiYing: LumiYing Terms of Service, none found (searched watermark, label, attribution, provenance, C2PA); undated. Checked
- OpenArt: OpenArt Terms of Service, Terms 4.5
Watermarks
; dated 30 July 2026. Checked - OpenArt: OpenArt Pricing, Pricing page plan cards (each paid card lists
Watermark-free
); undated. Checked - Pika: Pricing: Pika, Pricing page, Free card and comparison row
No watermark: Yes Yes Yes Yes
; undated. Checked - Pika: Help Center: Pika, FAQ
How can I get rid of the Pika watermark?
; undated. Checked - Adobe Firefly: Adobe Revolutionizes AI-Assisted Creativity with Firefly..., Press release, Firefly app section; dated 24 April 2025. Checked
- Leonardo.Ai: Generating Video on Leonardo.Ai, Help article FAQ, last question; dated 18 February 2026. Checked
- FLORA: FLORA Terms of Service, none found (searched watermark, C2PA, provenance, label); dated 17 November 2025. Checked
- ElevenLabs: ElevenLabs Pricing, Pricing page tooltip, Automatic dubbing; undated. Checked
- ElevenLabs: What is watermarking, and why is ElevenLabs using it?, Help centre, Audio Detector FAQ
What is watermarking, and why is ElevenLabs using it?
; undated. Checked - ElevenLabs: Audio detector, Audio detector docs,
Third-party models
; undated. Checked - Artlist: Artlist Terms of Use (Wayback capture), Terms of Use s.22 (Tools), not about AI outputs; dated 2 July 2026. Checked
- Google: Manage watermark settings in Gemini Apps (Gemini Apps Help), Manage watermark settings > Important; About invisible watermarks & Content Credentials; undated. Checked
- Google: Image generation with Gemini (Nano Banana): Gemini API, Gemini API docs > Image generation (stated twice, intro and limitations); dated 9 October 2026. Checked
- OpenAI: Content provenance: OpenAI API, Content provenance > What content provenance checks (table); undated. Checked
- Midjourney: Midjourney Image Verification, Midjourney Image Verification page; undated. Checked
- Ideogram: Ideogram Terms of Service, Terms of Service 6.2 Restrictions On Your Use of the Services (H); dated 14 August 2024. Checked
- Canva: Watermarks or Canva logos are on my design: Canva Help Center, Help Center: Watermarks or Canva logos are on my design; undated. Checked
- Canva: Canva AI Product Terms, AI Product Terms > Be a good human; dated 26 June 2026. Checked
- Meta: Meta Ad Creative Generative AI Terms, Ad Creative Generative AI Terms, 1. Ad Content > Misrepresentation; dated 6 May 2024. Checked
- Picsart: Picsart pricing, Pricing page translation bundle, key SOME_OF_THE_BENEFITS_OF_A_PICSART_GOLD_SUBSCRIPTION_INCLUDE; undated. Checked
- AdCreative.ai: AdCreative.ai homepage pricing table, Pricing tooltips on the homepage plan table (Unlimited Generations / Downloads); undated. Checked
- The Brief: The Brief pricing, no statement found; undated. Checked
- Pencil: Introducing Nano-banana (Gemini 2.5-flash Image), Blog: Introducing Nano-banana,
Enterprise Compliance
bullet; dated 23 September 2025. Checked - Omneky: Omneky Terms, no statement found; agreement sets no credit, watermark or AI-disclosure duty; dated 27 August 2025. Checked
- Predis.ai: Predis.ai pricing, Pricing page FAQ, HTML comment (not rendered); undated. Checked
- Blaze: Why Your Post Shows a Blaze Watermark, Help: Why Your Post Shows a Blaze Watermark; dated 20 March 2026. Checked
- Photoroom: New in product: August 2024, Inside Photoroom: New in product August 2024 > No more watermark for free users; dated 30 August 2024. Checked
- Claid.ai: Claid.ai pricing, no statement found; undated. Checked
- Flair.ai: Flair AI pricing, no statement found; undated. Checked
- Pixelcut: Pixelcut pricing, Pricing page, Free plan card; undated. Checked
- Pebblely: Pebblely pricing, no statement found; undated. Checked
- Meta: About AI info on ads created or edited with generative AI tools: Meta Business Help Center, ads surface; undated. Checked
- Meta: Labeling AI-Generated Images on Facebook, Instagram and Threads, organic surface; 6 February 2024 (page also shows 1 April 2025). Checked
- TikTok: TikTok ads policy: Misleading and false content (section ‘Edited media and AI-generated content (AIGC)’), ads surface; last updated April 2026. Checked
- TikTok: Partnering with our industry to advance AI transparency and literacy, organic surface; 9 May 2024. Checked
- YouTube: Disclosing use of GenAI content: YouTube Help, organic surface; undated. Checked
- Google Ads: Use AI content label settings and disclosures: Google Ads Help, ads surface; undated. Checked
- LinkedIn: Content credentials: LinkedIn Help, organic surface; Last updated: 7 days ago (read 12 Oct 2026, so about 5 Oct 2026). Checked
- EUR-Lex: EU AI Act, Article 50(2), machine-readable marking, Article 50(2); Article 113 (
It shall apply from 2 August 2026.
). Checked - EUR-Lex: EU Digital Omnibus on AI, Regulation (EU) 2026/1744, new Article 111(4) of the AI Act, Article 1, point (39)(b), adding Article 111(4); recital (38)
a transitional period of four months
; Article 1, point (20) replaces Article 50(7) on codes of practice; Article 4 entry into force. Checked - California Legislative Information: California AI Transparency Act (SB 942, chapter 291 of 2024), as amended by AB 853 (chapter 674 of 2025) and SB 1000 (chapter 861 of 2026), Business and Professions Code 22757.3(a) as amended by SB 1000 section 4; 22757.1(e) and (l); 22757.4; 22757.5(b); 22757.6. Checked
- California Legislative Information: California AI Transparency Act, large online platforms (Section 22757.3.1, added by AB 853, amended by AB 2713, chapter 856 of 2026), Business and Professions Code 22757.3.1(a), (c), (d), (e) as amended by AB 2713; definition 22757.1(i). Checked
- California Legislative Information: California SB 1050 (chapter 246 of 2026), Business and Professions Code 17610, synthetic performers in ads, Business and Professions Code 17610(b), (c); approved 16 Sep 2026. Checked
- New York State Assembly: New York S8420-A (chapter 617 of 2025), General Business Law 396-b, synthetic performers in ads, General Business Law 396-b(3); signed 11 Dec 2025 (
SIGNED CHAP.617
). Checked - Google DeepMind: Veo, Veo videos marked with SynthID; undated. Checked
- BytePlus: Service Specific Terms, 2.2, the invisible watermark on Seedance output via the API; updated 2026-10-08. Checked
- Kling AI: AI Content Detection, the imperceptible watermark and metadata in every output; undated. Checked
- TikTok: More ways to spot, shape and understand AI content, invisible watermarks on content made with TikTok tools and on uploads with C2PA Content Credentials; dated 19 November 2025. Checked
What changed on this page
- Updated for the 16 October 2026 price changes (Creatify’s new plan line-up, AdCreative.ai’s credit allowances and Predis.ai’s yearly prices): figures on this page follow.
- Page written: 63 tools and their plans, seven platform surfaces and six laws, each read at its official source.
Alexandre S.
Alexandre S. is a creative ad strategist and copywriter with ten years' experience in digital marketing across multiple channels.
Related pages
-
Data AI commercial use, tool by tool and plan by planCan a tool’s output run in your paid ads? Terms quoted with clause and date for each plan: allowed, conditional, barred or unclear. Checked 26 September 2026.
-
Data AI ad tool price tracker69 of 82 AI ad tools show a list price read on the vendor’s own page, last checked 16 October 2026, with the cost per clip or image where a rate allows.
-
Data How many ads are AI generated?Meta, TikTok, Google and Amazon publish no share of their ads made with AI. Six published figures, what each counts and who measured it, checked 2 October 2026.
