This Content Policy is part of the binding Terms of Service
This Content Policy, with an effective date of April 10, 2025, is incorporated into Hugging Face's Terms of Service, which the Policy describes as a binding agreement between you and Hugging Face. It says it sets out Hugging Face's actions to protect the Community on the Platform and gives information about the Content Hugging Face authorizes to be posted. It tells you to review Hugging Face's other guidelines, policies and documents on the Website, naming the Terms of Service, the Privacy Policy and the Code of Conduct; it describes the Code of Conduct as a general overview of Hugging Face's standards and of how it wishes the community would behave.
Supporting quotations
Effective Date: April 10, 2025
We created the following Content Policy to outline our actions to protect our Community on our Platform and to provide important information about the Content we authorize to be posted on our Platform.
This Policy is incorporated into our Terms of Service, which is a binding agreement between you and us.
You should carefully review all of our other guidelines, policies, and documents available on our Website
which gives a general overview and states our standards and how we wish the community would behave
Not stated here
The Terms of Service, Privacy Policy and Code of Conduct are linked but were not supplied with this document, so their contents are not analyzed here.
This Policy does not say whether the Code of Conduct is itself binding; it describes it as an overview of standards and of how Hugging Face wishes the community would behave.
Legal Rights & Agreement Rules · Context
Accessing or using the Platform counts as consenting to this Policy and other terms
By accessing or using the Platform, you consent to all of this Policy and to Hugging Face's other policies and terms. If you do not agree with any of them, the Policy asks you not to access or use the Platform.
Supporting quotations
By accessing or using our Platform, you consent to all of this Policy and our other policies and terms.
So, if you do not agree with any of those, please do not access or use our Platform.
Not stated here
The 'other policies and terms' are not listed exhaustively here, and those documents were not supplied with this analysis.
Legal Rights & Agreement Rules · Procedure
Policy changes take effect 10 days after posting; continued use means acceptance
Hugging Face may change or update this Policy from time to time. A change becomes effective 10 days after the updated Policy is posted on the Website, and if you keep using the Platform 10 days after that posting, the Policy treats that as your acceptance of the changes.
Supporting quotations
We may change or update this Policy from time to time.
Changes will be effective 10 days following posting the updated Policy on the Website.
If you continue using our Platform 10 days following such posting, that means you accept those changes.
Not stated here
This Policy mentions only posting the updated Policy on the Website; whether Hugging Face gives any other notice of changes (for example under the Terms of Service) is not stated here.
Legal Rights & Agreement Rules · Context
Undefined capitalized terms take their meaning from other Hugging Face documents
Capitalized terms that this Policy uses without defining have the meanings given in the Terms of Service, the Privacy Policy and all other policies on the Website. Terms used here but not defined here, such as 'User' and 'Website', therefore depend on documents that were not supplied with this analysis.
Supporting quotations
Capitalized terms used but not defined herein shall have the meaning assigned in our [Terms of Service](https://huggingface.co/terms-of-service), [Privacy Policy](https://huggingface.co/privacy), and all other policies available on our Website.
Not stated here
The Policy does not say which document controls if the Terms of Service, Privacy Policy and other policies define the same term differently.
Because the linked documents were not supplied, the meanings of terms such as 'User' and 'Website' are not verified here.
Content & IP Rights · Context
“Content” means any material on the Website or Hub, from anyone
Content is defined broadly as any material posted, displayed or accessed on the Website or Hub, including but not limited to code, data, text, graphics, images, usernames, applications or software, whether provided by you, by Hugging Face or by any third party. The Policy names two content types that Content 'may include': ML Artifacts (code and assets hosted as Hugging Face Repositories, including Models, Datasets and Spaces) and Community Content (content in the Community section of the Platform, including discussions, comments and usernames, plus related documentation such as READMEs, model cards, data cards, pull requests and merges). Both lists are open-ended.
Supporting quotations
“Content” refers to any material posted, displayed, or accessed on our Website or Hub, including but not limited to code, data, text, graphics, images, username, applications, or software you, we, or any third party provide or make available.
Content types may include:
“ML Artifacts”: Code and assets hosted as Hugging Face Repositories, including Models, Datasets, Spaces
“Community Content”: Content that can be found in the Community section of the Hugging Face Platform, including discussions, comments, and usernames, as well as related documentation such as READMEs, model cards, data cards, pull requests, and merges.
Qualifications & exceptions
The definition covers material that Hugging Face or third parties provide, not only what you upload.
Legal Rights & Agreement Rules · Context
“Community” includes Hugging Face staff; “Hugging Face” may act through others
Community means all Users of the Platform, including Hugging Face personnel. 'Hugging Face' means Hugging Face Inc., which may perform its obligations through its affiliates, directors, subsidiaries, contractors, licensors, officers, agents and/or employees. 'Team' or 'Hugging Face Team' means Hugging Face personnel.
Supporting quotations
“Community” refers to all Users of the Hugging Face Platform, including Hugging Face personnel.
“Hugging Face” refers to Hugging Face Inc., which may perform its obligations through its affiliates, directors, subsidiaries, contractors, licensors, officers, agents, and/or employees.
“Team” or “Hugging Face Team” refers to Hugging Face personnel.
The Hub is the huggingface.co platform for hosting versioned Repositories
The Platform (also called the 'Hugging Face Hub' or 'Hub') is the hosting platform at huggingface.co where Users can build, benchmark, share, version and deploy Repositories, which may include Models, Datasets and Machine Learning Applications. A Repository is a data structure containing all of the project files and the entire revision history. The Community Tab is a collaborative feature where the Community can discuss specific Repositories, including giving feedback, brainstorming ideas and opening pull requests. A Repository tag is a way to associate and present metadata of any kind related to a repository.
Supporting quotations
“Platform”, “Hugging Face Hub”, or “Hub” refers to the hosting platform available at huggingface.co where Users can build, benchmark, share, version, and deploy Repositories, which may include Models, Datasets, and Machine Learning Applications.
“Repository” refers to a data structure that contains all of the project files and the entire revision history.
“Community Tab” refers to a collaborative feature where the Community can discuss specific Repositories, including providing feedback, brainstorming ideas and opening pull requests for improvements.
“Repository tag” refers to a way to associate and present metadata of any kind related to a repository.
Content & IP Rights · Context
Repositories can be Public, Private, Gated or Disabled, with different access
A Repository may be: Public, meaning anyone on the internet can see it but only you or members of your organization can change it; Private, meaning only you or members of your organization can see and change it, and new users must join the maintaining organization to see it and access its Content; Gated, meaning the Repository and its Community Content are visible to everyone, but access to its ML artifacts (data, model weights) requires either accepting conditions in a click-through form or approval by the Repository maintainers; or Disabled, meaning access is blocked for all Community members except its owner.
Supporting quotations
A Repository may be:
“Public”: anyone on the internet can see it, but only you or members of your organization can make changes;
“Private”: only you or members of your organization can see and make changes to the Repository; New Users need to join the maintaining organization in order to both see the Repository and access its Content.
“Gated”: Gated Repositories and their Community Content are visible to everyone, but access to their ML artifacts (data, model weights) requires either accepting conditions in a click-through form or approval by the Repository maintainers.
“Disabled”: a Repository that has its access blocked to all Community members except its owner.
Not stated here
'Community' is defined to include Hugging Face personnel, so the Disabled definition read literally would also exclude staff; the Policy does not say whether Hugging Face personnel can still access Private or Disabled Repositories.
Clear violations of this Policy or the law are Restricted Content, subject to immediate action
Restricted Content includes clear violations of this Content Policy or of applicable laws, and the Policy says such Content is subject to immediate action; this section does not say what that action is. 'Content' is the Policy's broad defined term for any material posted, displayed or accessed on the Website or Hub, whoever provides it. The categories that follow are what Restricted Content 'may include', so the list is not exhaustive. If you see Content that violates the Policy, you are asked to report it, and the Hugging Face Team (Hugging Face personnel) addresses it on a case-by-case basis.
Supporting quotations
Restricted Content includes clear violations of our Content Policy or applicable laws, and is subject to immediate action.
Restricted Content may include:
If you see Content that violates this Policy, report it. It will be directly addressed by the Hugging Face Team on a case-by-case basis.
“Content” refers to any material posted, displayed, or accessed on our Website or Hub
“Team” or “Hugging Face Team” refers to Hugging Face personnel.
Qualifications & exceptions
'Immediate action' is stated for clear violations; the moderation section separately describes review of reported Content and a preference for collaborative solutions where possible.
Not stated here
Which jurisdiction's 'applicable laws' are meant is not stated.
Cross-references
contrast (model-proposed): Hugging Face prioritizes collaborative fixes with Repository owners where possible — contrast: Clear violations are described as subject to immediate action, while the moderation section says Hugging Face prioritizes collaborative solutions with the Repository owner whenever possible. The two can coexist (clear violations versus Content that modifications or guardrails can bring within the Policy), but the Policy does not say how it decides which path applies.
Content & IP Rights · Restriction or exception
Illegal Content and promotion of high-risk illegal activities are restricted
Restricted Content may include unlawful or illegal Content: all Content that violates any applicable law or regulation, and Content promoting high-risk illegal activities. The Policy's examples of such activities are weapons development, illegal substances, scams, gambling, pseudo-pharmaceuticals and plagiarism, followed by 'etc.', so the examples are not a complete list.
Supporting quotations
Unlawful or illegal Content
All Content that violates any applicable law or regulation.
Gambling and plagiarism are given as examples of 'high-risk illegal activities'; the Policy does not say whether promoting gambling or plagiarism that is lawful where it happens is covered.
Which jurisdiction's law or regulation is 'applicable' is not stated.
Content & IP Rights · Restriction or exception
Fraud, deception, unlawful financial schemes and unlicensed professional practice are restricted
Restricted Content may include fraudulent or malicious activities: unlawful, defamatory, fraudulent or intentionally deceptive Content (examples given are disinformation, phishing, scams and inauthentic behavior); Content that promotes or induces unlawful or fraudulent currencies, securities, investments or transactions; and Content that performs illegal or unlicensed professional practices, such as medical, legal or financial practice.
Content that promotes or induces unlawful or fraudulent currencies, securities, investments, or transactions.
Content that performs illegal or unlicensed professional practices (medical, legal, financial, etc.).
Qualifications & exceptions
Only the 'deceptive' item is limited to intentional deception; unlawful, defamatory and fraudulent Content are listed without an intent condition.
The financial item covers currencies, securities, investments or transactions that are unlawful or fraudulent, not financial Content generally.
Not stated here
The Policy does not explain when Content, such as a model or application, 'performs' an illegal or unlicensed professional practice.
Content & IP Rights · Restriction or exception
Harmful Content, hate speech, harassment and terrorist or violence-glorifying Content are restricted
Restricted Content may include harmful or abusive Content: Content that harms individuals or groups; Content promoting discrimination or hate speech (the Policy points to its Code of Conduct for this); Content involving harassment, bullying or demeaning behavior; and terrorist Content or Content that glorifies violence, suffering or humiliation.
Supporting quotations
Harmful or Abusive Content
Content that harms individuals or groups.
Content promoting discrimination or hate speech (see our Code of Conduct).
Content involving harassment, bullying, or demeaning behavior.
Terrorist Content or Content that glorifies violence, suffering, or humiliation.
Not stated here
The Code of Conduct referenced for discrimination and hate speech was not supplied with this document.
'Content that harms individuals or groups' is not further defined in this Policy.
Content & IP Rights · Restriction or exception
Non-consensual or harassing sexual Content, and any sexual Content involving minors, is restricted
Under harmful or abusive Content, Restricted Content may include sexual Content used for harassment or bullying, or created without explicit consent, and underage nudity or any sexual Content involving minors.
Supporting quotations
Harmful or Abusive Content
Sexual Content used for harassment, bullying, or created without explicit consent.
Underage nudity or any sexual Content involving minors.
Qualifications & exceptions
The first item is limited to sexual Content used for harassment or bullying or created without explicit consent; the minors item covers any sexual Content involving minors and any underage nudity, with no further condition.
Not stated here
Whose 'explicit consent' is required (for example, people depicted) is not stated.
Content & IP Rights · Restriction or exception
Exposing others' private information and infringing third-party IP are restricted
Restricted Content may include Content that violates the privacy of a third party, including but not limited to publishing other people's private information, such as a physical or email address, without their explicit permission; and Content that infringes the intellectual property rights of a third party.
Supporting quotations
Privacy and Intellectual Property (IP)
Content that violates the privacy of a third party, including but not limited to publishing others' private information, such as a physical or email address, without their explicit permission.
Content that infringes the intellectual property rights of a third party.
Qualifications & exceptions
The privacy example turns on publishing someone's private information without their explicit permission; the example list is open-ended ('including but not limited to').
Not stated here
What counts as 'private information' beyond the physical and email address examples is not defined in this Policy.
Security & Safeguards · Restriction or exception
Content built to break into or damage systems, or to spread or generate malware, is restricted
Restricted Content may include Content designed to disrupt, damage or gain unauthorized access to systems or devices, and Content that attempts to transmit or generate malicious code, with malware, trojans and viruses given as examples.
Supporting quotations
Platform Abuse, Security Violations and Spam
Content designed to disrupt, damage, or gain unauthorized access to systems or devices.
Content that attempts to transmit or generate malicious code (e.g., malware, trojans, viruses).
Qualifications & exceptions
The first item turns on Content being 'designed to' disrupt, damage or gain unauthorized access; the second covers Content that 'attempts to' transmit or generate malicious code.
Not stated here
The Policy does not say how the 'generate malicious code' item applies to general-purpose models or datasets that could be used to produce such code.
Content & IP Rights · Restriction or exception
Abusing Hugging Face services, bypassing restrictions, gaming metrics, cryptomining and spam are restricted
Restricted Content may include abuse of or interference with Hugging Face services, including using unauthorized bot APIs or remote management tools; hosting excessive or irrelevant data in repositories; using tools like Cloudflare Tunnel, TOR, proxies, VNC or Chrome Remote Server to bypass restrictions; incentivizing manipulation of Hub metrics, for example exchanging rewards for likes; and cryptomining practices. Spam is also listed, with examples of advertising products or services, excessive bulk activity, or disrupting the user experience.
Supporting quotations
Platform Abuse, Security Violations and Spam
Abuse or interference with Hugging Face services, including:
Using unauthorized bot APIs or remote management tools.
Hosting excessive or irrelevant data in repositories.
Using tools like Cloudflare Tunnel, TOR, proxies, VNC, Chrome Remote Server, etc., to bypass restrictions.
Incentivizing manipulation of Hugging Face Hub metrics (e.g., exchanging rewards for likes).
Cryptomining practices.
Spam (e.g., advertising products/services, excessive bulk activity, or disrupting user experience).
Qualifications & exceptions
The tools item is limited to using those tools to bypass restrictions; the list does not itself name other uses of them.
Not stated here
'Excessive or irrelevant data', the 'restrictions' being bypassed and what makes a bot API 'unauthorized' are not defined in this Policy.
Spam examples include 'advertising products/services' without a further limit; whether any advertising in Repositories or posts counts as spam is not stated.
Content & IP Rights · Permission or license
Hugging Face may also moderate Content outside the listed categories
Beyond the listed Restricted Content categories, Hugging Face says it may also moderate other types of Content in response to evolving challenges posed by advances in machine learning. In assessing such Content, it says it holds consent as a core value and aims for an approach that is thoughtful, adaptive and respectful of individual and community rights.
Supporting quotations
While the categories of Restricted Content above provide a clear framework, we may also moderate other types of Content in response to evolving challenges posed by advancements in Machine Learning.
As we assess such Content, we hold consent as a core value, ensuring our approach remains thoughtful, adaptive, and respectful of individual and community rights.
Qualifications & exceptions
This is a 'may': it reserves discretion rather than committing to moderate any particular type of Content.
Not stated here
No criteria, examples or notice process are given for these other types of Content.
Use & Processing · Company practice
The Hub shows everyone the same public content, without personalized recommendations
Hugging Face states that Users see the same public content on the Hub without personalized recommendations. Trending content is influenced by the number of likes in the past few days. Posts and updates appear from accounts a user actively follows, in strict chronological order, without ranking or algorithmic curation.
Supporting quotations
Users see the same public content on the Hub without personalized recommendations.
Trending content is influenced by the number of likes in the past few days.
Posts and updates appear from accounts users actively follow, displayed in strict chronological order without ranking or algorithmic curation.
Qualifications & exceptions
This statement sits at the end of the Restricted Content section but describes how content is displayed, not a content restriction.
Not stated here
The length of 'the past few days' and how likes are weighted for Trending are not specified.
Content & IP Rights · Procedure
Report Content you think violates this Policy in the Platform or by email
If you encounter any Content on the Platform that you believe violates this Policy, you can report it in two ways. In-Platform: click a button and choose the 'Report' option; for Repositories this opens a public report and notifies the Hugging Face Team, and for Posts and Comments it notifies the Team directly. By email: send a report to safety@huggingface.co. Separately, Australian residents can report illegal content to aus-online-safety@huggingface.co, which the Policy says lets Hugging Face inform them of the outcome of its investigation and any actions taken.
Supporting quotations
Any Content
If you encounter Content on our Platform that you believe violates this Policy, you can report it to us through the following channels:
In-Platform Reporting: Click the button and select the "Report" option.
For Repositories: Open a public report and notify the Hugging Face Team
For Posts and Comments: Notify the Hugging Face Team directly
Email: Send a report to safety@huggingface.co
To report illegal content, Australian residents can email aus-online-safety@huggingface.co. This enables Hugging Face to inform you of the outcome of our investigation and any actions taken in response to the report.
Qualifications & exceptions
A report on a Repository is opened as a public report; reports on Posts and Comments go directly to the Team.
Not stated here
The label or icon of the button to click is missing from the captured text, so the exact control is not identified here.
Apart from the Australian route, this section does not say whether reporters are told the outcome of their report.
Content & IP Rights · Restriction or exception
Reports are themselves Community Content, and abusive flagging is not tolerated
Reports you file are Community Content and are themselves subject to this Policy. Abusive uses of the flagging feature, including but not limited to spamming or harassment, will not be tolerated. A report on a Repository is opened as a public report.
Supporting quotations
Please note that reports are Community Content and are themselves subject to this Policy.
Abusive uses of the flagging feature, including but not limited to spamming or harassment, will not be tolerated.
For Repositories: Open a public report
Not stated here
The consequence of abusing the flagging feature is not specified beyond 'will not be tolerated'.
Content & IP Rights · Permission or license
Hugging Face may flag Content itself, including from automatic detection
In some situations the Hugging Face Team may flag Content to monitor requests or concerns raised through other channels, for example social media, automatic detection or concerns raised internally. Depending on the severity of the issue, Hugging Face may then take action as described in its moderation section.
Supporting quotations
In addition, in some situations, the Hugging Face Team may flag Content to monitor requests or concerns expressed through other channels (e.g., via social media, automatic detection, or raised internally).
Depending on the severity of the issue, we may take action as described in the Section below, “How we moderate Content”.
Qualifications & exceptions
Both the flagging and any follow-up action are stated as things Hugging Face 'may' do, depending on severity.
Not stated here
The Policy does not describe the automatic detection used or the criteria for flagging.
IP owners can send a takedown notice; Hugging Face reviews it and may disable the Content
If you believe Content on the website infringes your intellectual property rights, you can submit a takedown notice to dmca@huggingface.co; the Policy says this is a process it follows according to applicable law. You should include detailed and accurate information to support your claim, and by submitting a complaint you confirm that the information is truthful and that you will not intentionally submit false or misleading details. Hugging Face will review the notice to check that it meets legal requirements and is not abusive; if that is confirmed, it will inform the User who uploaded the allegedly infringing Content and disable that Content.
Supporting quotations
IP Infringement and DMCA takedown process
If you believe that any Content on our website infringes upon your intellectual property rights, you can submit a
to dmca@huggingface.co. This is a process we follow according to applicable law.
Please include detailed and accurate information to support your claim. By submitting a complaint, you confirm that all information provided is truthful and that you will not intentionally submit any false or misleading details.
We will review the takedown notice to ensure it meets legal requirements and isn’t abusive. If this is confirmed, we will inform the User who uploaded the allegedly infringing Content and disable it.
Qualifications & exceptions
Disabling follows only if the review confirms that the notice meets legal requirements and is not abusive.
Not stated here
The process is described for any intellectual property rights, while the heading names the DMCA and the linked sample notice is hosted on copyright.gov; whether non-copyright claims follow exactly this process is not stated.
The linked sample notice was not supplied with this document.
Uploaders can counter-notify; the claimant then has 14 U.S. business days to sue
If your Content is disabled after a takedown notice and you think there has been a mistake, you can challenge the decision by sending Hugging Face a counter-notification, again including detailed and accurate information to support your claims. Hugging Face will review the counter-notification to check that it meets legal requirements; if it does, Hugging Face will inform the original claimant, who will have 14 U.S. business days to take legal action to prevent the Content from being restored.
Supporting quotations
The original uploader can challenge this decision if they think there has been a mistake by submitting a
to us and similarly including detailed and accurate information to support their claims.
We will review the counter-notification to ensure it meets the legal requirements, and, if it does, we will inform the original claimant, who will have 14 U.S. business days to take legal action to prevent the Content from being restored.
Qualifications & exceptions
The 14 U.S. business days run for the original claimant after Hugging Face informs them of a counter-notification that meets the legal requirements.
Not stated here
The Policy implies but does not expressly state that the Content is restored if the claimant takes no legal action within 14 U.S. business days, and gives no restoration timing.
No address is given in this paragraph for counter-notifications beyond sending them 'to us'; the linked sample counter-notice was not supplied.
Content & IP Rights · Procedure
The Hugging Face Team reviews reported or flagged Content, iteratively and with owners where possible
If Content is reported or flagged as potentially problematic, the Hugging Face Team (defined as Hugging Face personnel) reviews it to determine whether it meets this Content Policy. The Policy describes this as an iterative approach to deciding whether, and under what conditions, the Content may be hosted on the Platform. Repository reports will, when possible, be addressed in collaboration with both the Repository owner and any concerned party as part of the decision.
Supporting quotations
If Content is reported or flagged as potentially problematic, the Hugging Face Team will review it to determine whether it meets the terms and conditions of this Content Policy.
This requires an iterative approach to determine whether and under what conditions it may be hosted on the Platform.
Repository Reports will, when possible, be addressed in collaboration with both the Repository owner and any concerned party, as part of our decision-making process.
“Team” or “Hugging Face Team” refers to Hugging Face personnel.
Qualifications & exceptions
Collaboration with the Repository owner and concerned parties is stated 'when possible', not in every case.
Not stated here
No timeframe for review is stated.
Content & IP Rights · Permission or license
Moderation actions range from requested changes to account suspension or termination
After receiving and reviewing a report, the Team will take action on the Content where appropriate. Possible actions include, but are not limited to: asking the relevant User for collaboration or modifications to the Content; unranking the Content; adding a Not for All Audiences (NFAA) tag; removing or disabling access to the Content; restricting interactions; and suspending or terminating the account.
Supporting quotations
After receiving and reviewing a report, our Team will take action on the Content where appropriate. These actions may include, but are not limited to:
Asking the relevant User for collaboration or modifications to the Content
Unranking the Content
Adding a Not for All Audiences (NFAA) Tag
Removing or Disabling access to the Content
Restricting Interactions
Account Suspension or Termination
Qualifications & exceptions
Action is taken 'where appropriate', and the list of actions is open-ended ('may include, but are not limited to').
Not stated here
This section does not say whether or how the affected User is notified of an action or told the reasons; the Terms of Service, which were not supplied, may address this.
'Unranking' and 'restricting interactions' are not further explained.
Content & IP Rights · Company practice
Hugging Face prioritizes collaborative fixes with Repository owners where possible
Hugging Face says it prioritizes collaborative solutions involving the Repository owner, for both ML Artifacts and Community Content, whenever possible, especially where modifications or additional guardrails can help the Content meet this Policy. It also says it offers moderation tools on the Platform, described as a regularly updated list, so the Community can play an active role in moderation.
Supporting quotations
We prioritize collaborative solutions for both ML Artifacts and Community Content that involve the owner of the Repository whenever possible, especially in cases where modifications or additional guardrails can help the Content meet this Policy.
We offer a range of useful tools on our Platform to support Community Content moderation. This regularly updated list allows the Community to play an active role in moderation efforts.
Qualifications & exceptions
The priority applies 'whenever possible'; it is not a promise that every case is handled collaboratively.
Content & IP Rights · User right or choice
Repository owners can edit, hide, tag, gate and close content in their Repository
Repository owners can: edit discussion and pull request titles; hide or edit a community comment; tag their Content as 'Not For All Audiences'; gate access to their Repository so that they manually review and approve or reject access to their ML artifacts (the Policy links documentation for gated Models and Datasets); and close, lock or delete discussions and pull requests.
Supporting quotations
Repository owners can:
Edit discussions and PRs titles
Hide or edit a community comment
Tag their Content as being “Not For All Audiences”
Gate the access to their Repository, allowing the owners to manually review and approve/reject access to their ML artifacts
Close, lock, or delete discussions and PRs
Edit their own comments
Qualifications & exceptions
The owner list says owners can hide or edit 'a community comment', while the author list limits authors to 'their own comments'.
Not stated here
The Policy calls these tools a regularly updated list, so the tools actually available may differ from those listed.
The linked gating documentation was not supplied.
Content & IP Rights · User right or choice
Authors of a discussion or pull request can edit its title, manage their own comments and close it
Authors of a discussion or a pull request can edit its title, edit their own comments, hide their own comments, and close the discussion or pull request.
Supporting quotations
Authors of a discussion or a pull request can:
Edit the discussion or PR title
Edit their own comments
Hide their own comments
Close the discussion or PR
Qualifications & exceptions
Authors' comment controls are limited to their own comments.
Content & IP Rights · User right or choice
Contest a decision to disable your Content or suspend or terminate your account by email
If you disagree with a moderation decision to disable your Content or to suspend or terminate your account, you can send a complaint to safety@huggingface.co. You are asked to include all arguments and evidence supporting your claims so that Hugging Face can reconsider its decision. Hugging Face says it will diligently review your complaint and respond to you.
Supporting quotations
If you disagree with a moderation decision to disable your Content or suspend or terminate your account, you can send us your complaint at safety@huggingface.co.
Please include all arguments and evidence you may have to support your claims so we can reconsider our decision.
We will diligently review your complaint and will respond to you.
Qualifications & exceptions
The complaint route is described for decisions to disable Content or to suspend or terminate an account; the Policy does not say whether other actions, such as unranking, an NFAA tag or restricted interactions, can be contested this way.
Not stated here
No deadline is stated for filing a complaint or for Hugging Face's response.
Cross-references
qualifies (model-proposed): Moderation actions range from requested changes to account suspension or termination — qualifies: The complaint route covers decisions to disable Content or suspend or terminate an account, which correspond to two of the listed moderation actions; the Policy states no complaint route for the other listed actions.
Legal Rights & Agreement Rules · User right or choice
If still unsatisfied, you can use a certified out-of-court dispute body under the DSA
If you remain unsatisfied after complaining, the Policy says that, in accordance with the Digital Services Act, you have the right to approach a certified out-of-court dispute settlement body (it links an EU Commission page about these bodies). Hugging Face commits to participating in that process in good faith, with the aim of reaching a resolution.
Supporting quotations
If you remain unsatisfied, and in accordance with the Digital Services Act, you have the right to approach a certified
We commit to participating in this process in good faith, with the aim of reaching a resolution.
Qualifications & exceptions
The right is framed 'in accordance with the Digital Services Act', the EU law the Policy also names in its EU contact section.
Not stated here
The Policy does not state who is eligible, for example whether users outside the EU can use this route.
Whether an outcome binds either party, and any costs, are not stated; the Digital Services Act itself is not supplied here.
Users can contact safety@huggingface.co about this Policy and moderation
For all communications about this Content Policy and how Hugging Face moderates Content, Users are asked to contact safety@huggingface.co. This is the same address the Policy gives for email reports of violating Content and for complaints about moderation decisions.
Supporting quotations
For our Users: For all communications regarding our Content Policy and how we deal with moderating Content at Hugging Face, please contact us at safety@huggingface.co.
Email: Send a report to safety@huggingface.co
you can send us your complaint at safety@huggingface.co
Legal Rights & Agreement Rules · Procedure
EU authorities contact legal@huggingface.co, in English, including for terrorist-content removal orders
Under a label reading 'For EU residents', the Policy gives a contact for EU Member State authorities, the EU Commission and the European Board for Digital Services, as indicated in the Digital Services Act: legal@huggingface.co, for all communications about the Content Policy and how Hugging Face moderates Content. This is also Hugging Face's point of contact for removal orders related to the Terrorist Content Online Regulation. All communications should be made in English.
Supporting quotations
For EU residents
For EU Member State authorities, the EU Commission, and the European Board for Digital Services as indicated in the Digital Services Act: For all communications regarding our Content Policy and how we deal with moderating Content at Hugging Face, please contact us at legal@huggingface.co.
This is also our point of contact for removal orders related to the Terrorist Content Online Regulation.
All communications should be made in English.
Qualifications & exceptions
The address is stated for the named EU authorities, not for individual users.
Not stated here
The label reads 'For EU residents', but the text beneath it is addressed to EU authorities and gives residents no separate contact; whether EU residents should use legal@huggingface.co or the general safety@huggingface.co address is unclear.
The Digital Services Act and the Terrorist Content Online Regulation are named but not supplied here.
Legal Rights & Agreement Rules · Context
Hugging Face SAS in Paris is the EU main establishment, giving France's Arcom authority under the DSA
The Policy names Hugging Face, SAS as its main establishment in the European Union: a French société par actions simplifiée à associé unique registered in the Paris Trade and Companies Register under number 822 168 043, with headquarters at 9 rue des Colonnes, 75002 Paris, France. It says this designation gives authority to the French Regulatory Authority for Audiovisual and Digital Communication (Arcom) under the Digital Services Act. The Policy also links a DSA Transparency Report 2025. Elsewhere it defines 'Hugging Face' as Hugging Face Inc., which may perform its obligations through affiliates and subsidiaries.
Supporting quotations
The main establishment in the European Union is Hugging Face, SAS, a French société par actions simplifiée à associé unique registered in the Paris Trade and Companies Register under the number 822 168 043, and whose headquarters are located on 9 rue des Colonnes, 75002 Paris, France.
The designation of this main establishment in the European Union gives authority to the French Regulatory Authority for Audiovisual and Digital Communication (Autorité de régulation de la communication audiovisuelle et numérique - Arcom) per the Digital Services Act.
Our DSA Transparency Report 2025 is accessible by clicking
“Hugging Face” refers to Hugging Face Inc., which may perform its obligations through its affiliates, directors, subsidiaries, contractors, licensors, officers, agents, and/or employees.
Not stated here
The Policy does not explain the relationship between Hugging Face Inc. and Hugging Face, SAS beyond the definition allowing Hugging Face Inc. to act through affiliates and subsidiaries.
The linked DSA Transparency Report 2025 was not supplied, so its contents are not analyzed here.
Content & IP Rights · Restriction or exception
Uploaders must take appropriate steps to minimise risk content could generate child exploitation or pro-terror material
In a part headed as an additional content policy in compliance with Australian Section 13(3)(a) of the DIS Standard (Australian Online Safety), the Policy says that before uploading any content to the Hub, particularly any ML Artifacts, you must take appropriate steps to minimise the risk that the content could be used to generate child sexual exploitation material or pro-terror material.
Supporting quotations
Additional content policy in compliance with Australian Section 13(3)(a) of the DIS Standard (Australian Online Safety)
Before uploading any content to the Hub, particularly any ML Artifacts, you must take appropriate steps to minimise the risk that the content could be used to generate child sexual exploitation material or pro-terror material.
Qualifications & exceptions
The duty is to take 'appropriate steps' to minimise the risk, with particular emphasis on ML Artifacts; the Policy does not say which steps are appropriate.
Not stated here
The part is headed as an Australian addition, but the sentence is addressed to 'you' without a residence limit; whether it applies only to users in Australia is not stated.
Section 13(3)(a) of the DIS Standard is an external Australian instrument that is not supplied; its requirements are not described here.
Content & IP Rights · Procedure
Australian residents can report illegal content to aus-online-safety@huggingface.co
Australian residents can report illegal content by emailing aus-online-safety@huggingface.co. The Policy says this enables Hugging Face to inform the reporter of the outcome of its investigation and any actions taken in response to the report.
Supporting quotations
To report illegal content, Australian residents can email aus-online-safety@huggingface.co.
This enables Hugging Face to inform you of the outcome of our investigation and any actions taken in response to the report.
Qualifications & exceptions
This route is described for Australian residents reporting illegal content; the general reporting channels cover any Content believed to violate the Policy.
Not stated here
No timeframe is stated for informing the reporter of the outcome.