.mcp

Q2.2-1 - Provide the applying entity’s self-certification document, signed by the CEO, President, CFO, and/or equivalent officer of the applying entity. If financial statements are provided by a Qualified Parent Entity (QPE) of the applying entity, the CEO, President, CFO, and/or equivalent officer of the QPE must co-sign the certification document. The self-certification document must represent and warrant: SC2.2-1.1 - As of the submission date of the application, the applying entity is a current registry operator or an affiliated entity of a current registry operator with one or more active Registry Agreements (RA). SC2.2-1.2 - The applying entity and/or a QPE will fund the startup and long-term operation of all of the applying entity’s current gTLDs and applied-for gTLD strings. SC2.2-1.3 - The applying entity and/or its officers are bound by law in its jurisdiction to represent financial statements accurately and the applying entity is in good standing in that jurisdiction.

Answered with a document. Attachments are not published by ICANN.

Q2.3-1 - Provide a document with a list of all of the applying entity’s current gTLDs and a list of all gTLDs for entities affiliated with the applying entity (if applicable).

Answered with a document. Attachments are not published by ICANN.

Q5.1-1 - Provide the applying entity’s self-certification document, signed by the CEO, President, CFO and/or equivalent officer of the applying entity. The self-certification document must represent and warrant: SC5.1-1.1 - The applying entity will appropriately protect confidentiality of data and prevent unauthorized access to data and services. SC5.1-1.2 - The applying entity will maintain a mature, appropriately funded and staffed security program, following a recognized, modern security framework based on risk management (such as the ISO27000 series, COBIT, HITRUST CSF, legally required security frameworks, or equivalent). The security program must be in place prior to delegation, and exist through at least the period of the registry agreement. SC5.1-1.3 - The applying entity is aware of and has designed its systems and business to comply with the relevant privacy and security regulations for all countries in which it operates.

Answered with a document. Attachments are not published by ICANN.

Q5.2-1 - Provide the applying entity’s self-certification document, signed by the CEO, President, CFO and/or equivalent officer of the applying entity. The self-certification document must represent and warrant: SC5.2-1.1 - The applying entity will, no later than delegation of the Top Level Domain (TLD), establish a dedicated abuse point of contact responsible for addressing matters requiring expedited attention and providing a timely response to abuse complaints concerning any name registered in the TLD. SC5.2-1.2 - The applying entity will, no later than delegation of the TLD, establish, publish, and provide to ICANN the location of a mechanism for members of the public to submit reports of abuse in accordance with the current obligations of the Base RA and any Consensus Policies. SC5.2-1.3 - The applying entity has developed proposed measures for removal of orphan glue records for names removed from the zone when provided with evidence in written form that the glue is present in connection with malicious conduct (see Specification 6). SC5.2-1.4 - The applying entity has or will have at time of delegation, established policies for handling complaints regarding abuse. Such policies are to be maintained and posted publicly so that anyone can review the policies via the Internet and any other means deemed appropriate by the applying entity. The applying entity’s policies at a minimum should contain appropriate confirmation of the receipt of the abuse report, the process of review of the report, and actions that will be taken if the applying entity confirms the report is legitimate. SC5.2-1.5 - The applying entity understands that DNS Abuse is Phishing, Malware, Botnets, Pharming and Spam (when used to deliver other forms of DNS Abuse). The applying entity understands and is prepared to contribute to the mitigation or disruption of DNS Abuse in domains in the TLD zone. SC5.2-1.6 - The applying entity’s abuse response capabilities are resourced appropriately to ensure a timely and adequate investigation and response to reports of DNS Abuse. This includes capabilities to receive and evaluate evidence of DNS Abuse in reports, and to take action to stop or disrupt the DNS Abuse. SC5.2-1.7 - The applying entity is prepared to conduct periodic scans of its zone to identify if domains are being used to perpetrate DNS Abuse, and to maintain statistical reports of the scans, the findings, and actions taken.

Answered with a document. Attachments are not published by ICANN.

Script of String

Latin

As per Section 3(d) of Specification 11 of the Base Registry Agreement, a registry operator of a “generic string” may not impose eligibility criteria for registering names in the TLD that limit registrations exclusively to a single person or entity and/or that person’s or entity’s “Affiliates” (as defined in Section 2.9(c) of the Registry Agreement). “Generic String” means a string consisting of a word or term that denominates or describes a general class of goods, services, groups, organizations or things, as opposed to distinguishing a specific brand of goods, services, groups, organizations or things from those of others. Confirm that the applied-for string is not a “generic string” in which the applying entity intends to limit registrations exclusively to a single person or entity.

true

How would you categorize your community?

The Model Context Protocol (MCP) community is best categorised as a highly technical, activity-based, open source software developer and industry-wide technology ecosystem. The community is united by a shared, active pursuit: the collaborative development, deployment, and standardisation of the open Model Context Protocol.

The community is clearly delineated and comprises four main, verifiable categories of members:

1) Project Contributors: Developers, programmers, and software engineers with a documented history of technical contributions to the official Model Context Protocol open source project. These contributions include code commits, pull requests, technical documentation, or core specification development, all of which are public and verifiable via official GitHub repositories.

2) MCP Server Operators: Individuals and organisations that manage, host, and operate active MCP servers. These operators demonstrate strict technical compliance with the official, current MCP specification.

3) Registry-Listed Entities: Registered owners or authorised representatives of active MCP servers and developer tools that are officially catalogued in the public MCP Registry maintained by the Model Context Protocol project.

4) Agentic AI Foundation (AAIF) Members: Members in good standing of the AAIF - an industry consortium established to define the frameworks, standards, and collaborations that will guide agentic AI forward. These members are subject to the AAIF Charter and other AAIF policies.

By operating within these distinct, activity-based categories, community members collaborate to establish a secure, semantic, and highly interoperable namespace.

What is the applying entity's connection to the community?

The applying entity, Charleston Road Registry Inc. (CRR), is an American company wholly owned by Google LLC, whose ultimate parent company is Alphabet Inc. Google is a core technical stakeholder, active co-developer, and major industry sponsor within the Model Context Protocol (MCP) community. It is a platinum member of the AAIF, the organization that supports the Model Context Protocol project. Rather than acting as a detached registry service provider, CRR’s connection to the community is organic, direct, and technically integrated through Google's substantial contributions to the ecosystem.

CRR’s connection to the MCP community is built upon three primary pillars:

Technical and Administrative Endorsement: CRR is applying to operate the applied-for gTLD with the explicit, formal written support of the Agentic AI Foundation (AAIF), a directed fund of The Linux Foundation (LF), along with LF Projects, LLC (LFP) - the neutral administrative host and supporting organization for the Model Context Protocol, formally organized under Model Context Protocol a Series of LF Projects, LLC (MCP), and Anthropic - the creator of the Model Context Protocol. This collaborative alignment ensures that the operation of the applied-for gTLD remains strictly in the best interests of the open source developer ecosystem.

Operational Rationale (Security and AI Safety): The technical purpose of this application is to solve a critical infrastructure challenge within the AI development community. Currently, AI models locate MCP servers via traditional, complex, and unstandardised web URLs. CRR will leverage Google Registry’s infrastructure to establish a preloaded HSTS namespace for the applied-for gTLD. This allows host machines and AI agents to utilise semantic, cleanly structured lookups (such as weather.mcp or postgres.mcp) to verify that the host complies with strict AI safety, cryptographic certification, and machine-to-machine transport-layer authentication protocols.

Enforcement of Community Standards: In complete alignment with community leadership, CRR has committed to a binding Specification 12 Community Registration Policy. CRR will neutrally administer the applied-for gTLD to restrict registrations exclusively to verified open source contributors, active server operators, listed registry entities, and members in good standing of the Agentic AI Foundation (AAIF), preventing the namespace from being hijacked, warehoused, or abused.

How is the community organized? Are there one or multiple organizations ("organizing body") that represent or administer the community?

The Model Context Protocol (MCP) community operates under a structured, and neutral multi-stakeholder governance model administered by recognized industry-leading bodies and open source foundations:

Neutral Governance: Agentic AI Foundation (AAIF), a directed fund of The Linux Foundation (LF), along with LF Projects, LLC (LFP) are the neutral administrative host and supporting organization for the Model Context Protocol, formally organized under Model Context Protocol a Series of LF Projects, LLC (MCP). As the official governing host of the Model Context Protocol project since it was contributed by Anthropic in December 2025, MCP provides open source administrative oversight, and is backed by AAIF, which supports the MCP community globally and manages community outreach and events. It ensures the protocol remains a neutral, open specifications unencumbered by proprietary interests.

Anthropic (Protocol Creator): Anthropic created the Model Context Protocol and released it as an open standard in November 2024. In December 2025, Anthropic contributed MCP to the Agentic AI Foundation, a directed fund under the Linux Foundation, and continues to contribute to MCP through the AAIF.

The coexistence of these highly established, global organising bodies provides an independent administrative structure, formalises community participation, and guarantees that the community's pursuits are enduring, neutral, and sustainable over time.

LF Projects, LLC, AAIF, and Anthropic have provided written support of this Community application.

Does the community have defined membership requirements, such as registration, licensing, or use of specific communication? Or, do community members self-identify as part of the community?

The Model Context Protocol (MCP) community comprises both formal membership structures and structured, objectively verifiable self-identification pathways. Because the community is built around an open-source, multi-stakeholder artificial intelligence specification, participation is open to any individual or organisation that actively contributes to or implements the protocol, yet remains clearly delineated through public and verifiable technical benchmarks.

1. Structured Self-Identification via Technical Benchmarks - The vast majority of the community’s active members - namely software developers, open source contributors, and system operators - self-identify as part of the community through their voluntary participation in the protocol's development and implementation. This self-identification is not merely subjective, but is verified through objective, publicly reviewable mechanisms:

a) Open Source Contribution: Developers "join" the project contributor segment by submitting code commits, pull requests, bug reports, or core specification documentation to the official MCP open source repositories. These contributions are publicly tracked and verifiable via GitHub or other official project code hosting platforms.

b) Protocol Implementation and Server Operation: Software engineers and organisations self-identify by developing and operating active, compliant MCP servers. To be recognised as an eligible registrant in the applied-for gTLD, these operators must demonstrate that their servers comply with the official, current MCP specification.

c) Ecosystem Registry Listing: Individuals and entities can register their MCP-compliant servers, developer tools, or AI integrations in the official, public MCP Registry maintained by the Model Context Protocol project. Listing on this neutral registry provides a transparent record of community involvement.

2. Formal Membership Structures - In addition to self-identified technical participants, the community is anchored by formal membership organisations that provide administrative, legal, and operational oversight:

Agentic AI Foundation (AAIF): The AAIF, a directed fund of The Linux Foundation, is a structured, formal industry association dedicated to safe AI-to-machine interface standardisation. Joining the AAIF requires signing a membership agreement, which includes a commitment to follow the organization’s formal charter, adhering to established code-of-conduct guidelines and other policies, and paying dues, which establishes a legally-defined category of community members in good standing. AAIF, along with LF Projects, LLC (LFP) are the neutral administrative host and supporting organization for the MCP community (since its contribution in 2025), formally organized under Model Context Protocol a Series of LF Projects, LLC.

Where is the community located?

Located globally, primarily within decentralised digital spaces such as GitHub, developer forums, and open source communication channels, as well as the global office locations of contributing technology organisations and software developers.

What is the estimated size of the community? This should take into account any regions listed in Question 138.

MCP community size measured by various metrics: SDK downloads exceed 170M/month and 1B/year. The project has 190k+ GitHub stars and 1,300+ quarterly contributors. The April 2026 North America Dev Summit drew 1,100 attendees representing 457 organisations.

What portion of the community do any organizing bodies represent or administer to?

While Model Context Protocol a Series of LF Projects, LLC and the Agentic AI Foundation (AAIF) administers to the entirety of the MCP community, its open source development is intentionally decentralised, allowing anyone globally to participate.

Do the organizing bodies demonstrate active and consistent efforts to engage and connect with the identified community and its members?

Anthropic established the open MCP specification, and the community (including Google) has since contributed to its evolution.

The organising bodies administering the Model Context Protocol (MCP) community - specifically LF Projects, LLC (LFP) and the Agentic AI Foundation (AAIF) - demonstrate active, consistent, and documented practices to engage and connect with community members. Across the two years leading up to the submission of this application (2024-2026), these bodies have led ongoing, multi-stakeholder initiatives categorised into four core areas:

1) Offering Support: Organising bodies provide continuous technical support, maintaining open source Software Development Toolkits (SDKs) in Python, TypeScript, and Java. Engineers from Anthropic, Google, and other member organizations contribute to the project’s public open-source repositories.

2) Sharing Information: Clear, open communication is maintained through a public registry of verified MCP servers and tools, which serves as a central discovery directory. Complete technical specifications are publicly available on open-access websites, supplemented by technical blogs, architectural guides, and active developer forums.

3) Responding to Specific Needs: In response to developer demands for secure, standardised context integration pathways, Anthropic released MCP as an open standard, and the community, including Google, has contributed to its development. To address transport-layer security concerns and prevent the malicious hijacking of AI-to-machine endpoints, the Registry Operator will preload the applied-for gTLD in the HTTP Strict Transport Security (HSTS) list.

4) Fostering and Strengthening Relationships The community is built on neutral, collaborative relationships governed by open source principles. Since the Model Context Protocol project joined the LF in 2025, the community’s governance has transitioned to a neutral, multi-stakeholder model, ensuring that the protocol's development remains unencumbered by proprietary interests. The organising bodies coordinate development outreach, host technical workshops and conferences, and collaborate directly with community members to promote cross-industry interoperability and safe AI interface standardisation.

What is the role of the applying entity in the engagement efforts listed in Question 141?

The applying entity, Charleston Road Registry Inc. (CRR), as a wholly owned subsidiary of Google LLC, plays an active role in the community engagement efforts. Rather than acting as a detached registry service provider, CRR and Google will coordinate directly with the primary organising bodies of the Model Context Protocol (MCP) community to support adoption, security, and developer outreach.

CRR's and Google’s active role in the community engagement pillars includes:

1) Offering Support: Google's engineers play a direct role in technical engagement, contributing to MCP core specifications. Google's engineers actively participate in public GitHub repositories, resolving technical bugs, reviewing pull requests, and providing debugging support.

2) Sharing Information: Google drives broad community awareness by publishing technical guides, architectural best practices, and developer blogs on Google platforms. These resources accelerate adoption and standardise semantic interface integrations.

3) Responding to Specific Needs: Google provides an ecosystem of cloud-hosted, fully-managed MCP servers that safely connect AI agents to real-time data and tools without requiring custom integrations. Further, to address developer demands for secure, reliable transport-layer paths, CRR will preload the applied-for gTLD in the HTTP Strict Transport Security (HSTS) list, ensuring secure machine-to-machine connections.

4) Fostering Relationships: Google provides leadership for the MCP community. It is a platinum member of AAIF and has a representative on the Governing Board.

Google’s extensive technical resources and direct sponsorship guarantee that the community's pursuits are robustly supported, highly visible, and sustainable over the operational lifetime of the applied-for gTLD.

Are community members aware of the identified community and each other?

Yes, members of the Model Context Protocol (MCP) community demonstrate a high level of mutual awareness, recognising both the overarching community and the distinct member segments that form the ecosystem. Because the protocol is an open specification, this mutual awareness is preserved and documented through shared public infrastructure, developer indices, and collaborative open source governance.

1) Structural Alignment and Shared Infrastructure - Rather than operating as an unstructured group, members of the MCP community interact and identify each other through standardised technical benchmarks:

a) Open Source Collaboration: Software developers and contributors actively collaborate on core specifications and toolkits within the official MCP open source project repositories, tracking contributions publicly via GitHub.

b) Central Registry Indexing: The community maintains an official, public MCP Registry. Developers and server operators register their compliant implementations in this central registry, allowing users and AI agents to discover active tools.

c) Semantic Interoperability: This mutual technical awareness is demonstrated in practice through the deployment of protocol implementations. For example, AI agents and developers can easily identify and connect with specific compliant tools by looking up semantic, cleanly structured hosts.

2) Shared Governance and Strategic Alignment - The transition of the protocol’s administration has formalised and solidified mutual awareness across both commercial and open source stakeholders:

a) Neutral Governance Umbrella: Following Anthropic’s creation of the Model Context Protocol in 2024, it moved the project to LF Projects, LLC in 2025 to be supported by Agentic AI Foundation. Agentic AI Foundation (AAIF), a directed fund of The Linux Foundation (LF), along with LF Projects, LLC (LFP) are the neutral administrative host and supporting organization for the Model Context Protocol, formally organized under Model Context Protocol a Series of LF Projects, LLC (MCP) Operating under LF's neutral administrative framework provides the entire ecosystem with a unified, sustainable structure.

b) Corporate and Standardisation Coordination: Industry-wide alignment is further maintained through coordination between leading technical entities and standardisation bodies.

Through these active registry databases, shared GitHub repositories, and multi-stakeholder governance frameworks, members of the MCP community remain highly integrated, technically aligned, and deeply aware of each other’s active contributions.

Are community members aware of the applying entity and its intention to apply for a community gTLD?

The Agentic AI Foundation, and related entities, The Linux Foundation and LF Projects, LLC as representatives and governing body of the MCP Community are aware of and support the applying entity, Charleston Road Registry Inc. (CRR) and its intention to apply for the applied-for gTLD as a Community gTLD. Anthropic is similarly aware and supportive. Evidence of the community members’ awareness is evidenced by the endorsement letters included in this application.

This support will be published online following ICANN’s reveal of all gTLD applications.

Was there an established presence of the identified community prior to the opening of the application submission period?

Yes

Answered with a document. Attachments are not published by ICANN.

Are individuals and groups outside of the identified community aware of the existence of the identified community?

Yes, individuals and groups outside the immediate Model Context Protocol (MCP) developer community demonstrate a clear, documented awareness of the community’s existence and its active technical pursuits. As an open protocol designed to solve critical artificial intelligence connectivity challenges, the community’s established presence is recognised globally by technology organisations, enterprise networks, and standardisation bodies.

To demonstrate this established presence and external awareness, the community’s organising bodies have documented active practices across five core pillars during the two years leading up to the submission of this application (2024–2026):

1) Media and Public Information regarding the Community - Widespread technical and mainstream media coverage has documented the rapid rise and adoption of the Model Context Protocol since its creation by Anthropic in 2024. Widespread global coverage occurred in late 2025 when Anthropic formally contributed the protocol to the Linux Foundation and LF Projects, LLC (together "LF") to establish neutral, open source governance under the Agentic AI Foundation. This landmark contribution was reported by global technology journalists, open source news outlets, and industry commentators, highlighting to the general public that the protocol has become the standard for artificial intelligence data integration. This broad public reporting demonstrates that the community’s activities are highly visible to those outside the immediate development base.

2) Discussion of the Community in Technical Fora - Beyond core developer channels, the protocol is discussed extensively in broader technology, enterprise software, and artificial intelligence forums. It is recognised by enterprise architecture committees, database developers, and web standard groups as a revolutionary framework that replaces messy, proprietary APIs with semantic, cleanly structured lookups. These public discussions demonstrate that IT professionals and corporate decision-makers outside the core development group are highly aware of the community's existence and technical progress.

3) Evidence of Prior Chartering and Organisation - The formal chartering and transition of the protocol to AAIF, LF Projects, LLC, and Model Context Protocol a Series of LF Projects, LLC in 2025 provides concrete, legal proof of the community's established presence. This formal governance structure under a universally recognised, neutral open source foundation ensures that individuals and corporate groups outside the direct developer base can reliably participate in, contribute to, and adopt the protocol under established open source legal frameworks.

4) Technical and Scientific Contributions to a Larger Population - The protocol represents a major scientific and technical contribution to the broader digital society by solving a critical security and interoperability bottleneck. By standardising how AI agents connect safely to data sources, the community has significantly advanced global AI safety, machine-to-machine transport-layer authentication, and public standards for secure artificial intelligence deployment, benefiting the wider Internet population who rely on secure AI integrations.

Are the pursuits of the identified community enduring and sustainable?

Yes, the pursuits of the Model Context Protocol (MCP) community are enduring, sustainable, and structurally built for long-term growth. This longevity is demonstrated by the protocol’s rapid industry-wide adoption since its creation by Anthropic in 2024, and its transition to the neutral administrative governance of Model Context Protocol a Series of LF Projects, LLC and its supporting organization, the Agentic AI Foundation (AAIF).

To demonstrate the longevity and sustainability of the community's pursuits, the following practices have occurred within the two years leading up to the submission of this application:

1) Established and Enduring Open Source Governance: Following its initial development in 2024, the protocol transitioned to the neutral, open source stewardship of AAIF in 2025. This transition established a stable, permanent administrative framework designed to scale and add long-term value to the open source developer ecosystem, ensuring the protocol remains free from proprietary lock-in.

2) Recurring and Scheduled Activities: The community participates in continuous, structured collaborations, including ongoing codebase commits, regular software developer kit (SDK) version releases in Python, TypeScript, and Java, and public technical reviews of protocol specifications. This persistent development cycle demonstrates a long-standing technical tradition that will endure over the operational lifetime of the applied-for gTLD.

3) Widespread Enterprise and Industry Endorsement: The longevity of these pursuits is backed by leading technology companies and standardisation bodies.

This multi-stakeholder support ensures that the MCP community is forward-looking and highly sustainable. Supporting documentation and articles illustrating LF’s established track record of scaling open source communities, alongside the official MCP community charters, are attached to this application to substantiate these pursuits:

The Model Context Protocol Project Governance: https://modelcontextprotocol.io/community/governance

Agentic AI Foundation Charter: https://github.com/aaif/foundation/blob/main/charter.md

The Linux Foundation Annual Report: https://www.linuxfoundation.org/hubfs/Publications/2025%20Linux%20Foundation%20Annual%20Report_122225a_lr.pdf

Answered with a document. Attachments are not published by ICANN.

Does the string match the name of the identified community?

Yes, “MCP” is a commonly used acronym for the Model Context Protocol.

Will the general public instinctively think of the community when thinking of the applied-for string?

In today’s agentic online environment, the acronym “MCP” is strongly connected to Model Context Protocol. It is the first result from a Google search of “mcp” (see below). The acronym and protocol itself has become and will continue to be the backbone of modern AI agents, secure enterprise AI systems, and agentic workflows on the DNS.

CRR acknowledges that “mcp” has historically been used as an acronym for other terms, including: Master Control Program (1982 movie Tron), Metacarpophalangeal, Master of Clinical Psychology, Master Cleaning Program, Minor Compromise Petition, Metric Capital Partners, and Microsoft Certified Professional although none of these terms constitute a community as anticipated in the Applicant Guidebook.

Answered with a document. Attachments are not published by ICANN.

Are you proposing to include one or more Community Registration Policies in the Registry Agreement (RA) that are unique to the applying entity's applied-for community gTLD?

Yes

Please state a specific Community Registration Policy with respect to registration eligibility for community members.

This Registration Policy (the "Policy") sets forth the criteria and requirements for the registration and use of domain names within the .mcp top-level domain (TLD). .mcp is a community-based TLD which will solely serve the Model Context Protocol (MCP) community, including contributors, developers, and operators of MCP-compliant systems. The purpose of .mcp is to provide a dedicated, secure, and recognizable namespace that fosters the growth and interoperability of the MCP ecosystem. The Policy, as adopted by the .MCP Policy Council, including the Community Protection Commitments, will be enforced by the Registry Operator with the assistance of the .MCP Policy Council and may be amended only with the consent of the .MCP Policy Council.

Community Protection Commitments

Eligibility Only the following types of individuals and organizations are eligible to register domain(s) in .mcp:

1) Project Contributors: documented history of contributions to the official Model Context Protocol open source project, including but not limited to code commits, documentation, extension development or core specification development, verifiable via GitHub or other official project repositories. 2) MCP Server Operators: operation of an active MCP Server that demonstrates compliance with the current MCP specification. 3) MCP Registry Listing: registered owner and/or authorized representative of a server listed in the official MCP registry maintained by the Model Context Protocol project. 4) AAIF Members: members in good standing of the Agentic AI Foundation (AAIF), subject to adherence to the AAIF charter and community compliance standards.

State a specific Community Registration Policy with respect to name selection criteria or rules for the applied-for string.

Naming Selection Registry Operator shall restrict the selection of second-level domain names such that all registered domain names relate to the registrant's specific Model Context Protocol (MCP) implementation, their MCP service name, or their recognised brand within the developer community. Acceptable domain names must communicate that the registrant's business, service, or API is natively designed to integrate with Large Language Models (LLMs) and agentic artificial intelligence.

State a specific Community Registration Policy with respect to an additional commitment besides registration eligibility for community members and naming selection criteria or rules for the applied-for string.

Other Commitments

1) Community Alignment: domain names in the .mcp are strictly limited to hosting MCP related content, services, APIs, or documentation directly related to the MCP ecosystem. General-purpose use unrelated to the MCP protocol, use, or its development is prohibited. 2) Trademark Usage: Registrants must comply with the MCP trademark policy. Use of domains in .mcp must not imply official endorsement by the Model Context Protocol technical project, AAIF, LF Projects, The Linux Foundation, or Anthropic unless such endorsement exists. 3) Security Obligations: Registry Operator shall preload .mcp in the HTTP Strict Transport Security (HSTS) list to enforce secure HTTPS-only connections across all active hosts.

Publication Registry Operator shall implement and publish this Policy on its website no later than the date on which the TLD is delegated in the DNS.

Registrar Obligations Registry Operator shall include requirements in its Registry-Registrar Agreement for registrars offering the .mcp TLD to (i) implement verification processes as approved by and made available by Registry Operator inline with this Policy and (ii) bind registrants to this Policy, that specifically reserve the rights of Registry Operator to suspend, delete or take other action on domain names in .mcp based on violation of this Policy.

Explain the rationale for any limitations to the Community Registration Policy proposed by the applying entity in Questions 151-153.

Not Applicable

Explain how the proposed Community Registration Policies of the applying entity meets the Registry Commitments Evaluation criteria 4 and 5?

The proposed Community Registration Policies for the applied-for gTLD satisfy all requirements under Registry Commitments Evaluation (RCE) Criteria 4 and 5.

1. Compatibility with Criterion 4 (No Duplication or Contradiction of Laws and Agreements) - The eligibility and name selection policies proposed for the applied-for gTLD do not duplicate, nor are they contrary to, applicable national laws, ICANN agreements, consensus policies, or temporary policies.

No Duplication: These policies are highly specific to the technical and operational parameters of the Model Context Protocol (MCP) open source standard. They do not duplicate general registration obligations already governed by standard ICANN Consensus Policies or the Base Registry Agreement.

No Contradiction: All criteria rely on objective, publicly verifiable developer and operational benchmarks - such as documented contribution history on GitHub repositories or the active operation of an MCP-compliant server. These rules do not conflict with local jurisdictions or data protection regulations (such as GDPR). Furthermore, they do not commit registry operators or registrars to actions that violate the Registrar Accreditation Agreement (RAA) or any applicable Consensus Policies.

2. Compatibility with Criterion 5 (Compatibility with ICANN's Bylaws and Content Neutrality) - The proposed policies are fully compatible with ICANN's Bylaws, specifically aligning with ICANN’s core technical mission to ensure the stable and secure operation of the Domain Name System.

Strict Content Neutrality: Under its Bylaws, ICANN is prohibited from regulating or enforcing content-based restrictions on the Internet. In complete alignment with ICANN Board directives on content neutrality, these policies are strictly non-content-restrictive.

Procedural and Eligibility Focus: Rather than regulating the substance of speech or the messages carried by services in the applied-for gTLD, the policies focus entirely on operational, procedural, and technical eligibility (who can register based on verifiable community standing) and structural naming rules (domain names must relate to the registrant's specific MCP implementation, brand, or service).

No Content Policing: Once an eligible community participant (such as an open source contributor or active server operator) is verified and registers a domain, they are free to deploy their compliant tools and services without any content policing from the registry. Consequently, ICANN will never be put in the position of enforcing content-based restrictions or passing judgment on speech.

3. No Additional Registry Services Required - The execution and enforcement of these Community Registration Policies utilize standardised, automated verification workflows. They are processed entirely within the capabilities of the pre-evaluated Main Registry Service Provider (RSP) and do not require the operation or evaluation of any new or additional technical Registry Services.

From where does the applying entity have the support to run the applied-for string on behalf of the identified community?

Charleston Road Registry Inc. (CRR) has secured the formal, written endorsements of the primary organising, governing, and technical bodies that represent and administer the Model Context Protocol (MCP) community. These endorsements demonstrate that a substantial majority of the community’s leadership fully supports CRR’s application to operate the applied-for gTLD for the benefit of the ecosystem.

The supporting entities include:

Agentic AI Foundation (Neutral Governing Body), a directed fund of The Linux Foundation and LF Projects, LLC (Trademark Owner): As the neutral administrative host and supporting organization for the Model Context Protocol project since December 2025, and the owner of the Model Context Protocol project’s trademarks, AAIF and LF Projects have both formally endorsed CRR’s application. They govern and represent the global open source community to ensure that the ongoing development and maintenance of the protocol remains open and neutral.

Anthropic (Protocol Creator): Anthropic, which created the Model Context Protocol and released it as an open standard in 2024, has provided a letter of support for CRR’s application.

These organisations represent the absolute majority of the community's leadership, developers, and active users. Operating the applied-for gTLD under the proposed community registration policies ensures a secure, standardised, and highly trusted namespace for the Model Context Protocol ecosystem.

Answered with a document. Attachments are not published by ICANN.

Is there any opposition to the applying entity, application, or applied-for string that the applying entity is aware of? If yes, please explain.

Charleston Road Registry Inc. (CRR) is not aware of any active, formal, or relevant opposition to this application or to the applied-for gTLD string. CRR is aware that the acronym ‘MCP’ is used by unrelated third parties in other fields. CRR is not aware that any such party intends to apply for the string or to oppose this application.

By submitting this Application, the applying entity confirms that it is submitting this Application with a good faith (“bona fide”) intent to operate the gTLD for which it has applied, and that the applying entity has read and understands the provisions of Section 5.2.3.1 Prohibited Communications and Activities of the Applicant Guidebook regarding the New gTLD Program rules prohibiting certain communications and activities to prevent parties from privately resolving string contention among themselves.

true

By submitting this Application, the applying entity confirms that it has read and understands the provisions of Section 5.2.3.1 Prohibited Communications and Activities of the Applicant Guidebook regarding the New gTLD Program rules prohibiting certain communications and activities to prevent parties from privately resolving string contention among themselves.

true