The Connected Apps Platform allows you to make your Connected Apps available on
our Connected Apps Registry. Connected Apps are apps that extend the
capabilities of Google's generative AI products or services that are registered
in our Connected Apps Registry. A connected app may be enabled using the
technical integration method(s) or interface(s) described in our documentation,
such as model context protocol (MCP) servers and Android MCP implemented via
AppFunctions.

The policies listed below apply to you and any Connected Apps you choose to
register with our platform. As a connected app developer, you play an important
role in maintaining a positive experience for everyone using Google products. We
need to curb abuses that threaten our ability to provide these services, and we
require that developers abide by the policies below to help us achieve this
goal. If a policy violation has been identified, we will take action, including
rejecting the connected app at submission, suspending the connected app,
removing the connected app from our platform, or suspending and/or terminating
your developer account.

Check back from time to time as these policies are occasionally updated. It is
your responsibility to monitor and ensure your compliance with these policies on
a regular basis. If, at any time, you cannot meet the requirements of our
policies (or if there is a significant risk that you will not be able to meet
them), stop using our services and reach out to us.

> [!NOTE]
> **Note:** These policies apply to all code you implement, including third-party code your connected app implements or hosts.

## Connected app quality

Connected app quality is important for both users and developers. High-quality
Connected Apps deliver value to users and are designed for safety. A
high-quality connected app is one that is free of critical bugs, crashes, and
other technical issues; transparent to the user; and provides a safe and
high-value user experience.

Connected Apps must

- perform as described without causing model confusion or task failure.
- deliver tangible functionality and be reliable enough to handle valid requests without crashing or hanging.
- provide a stable, responsive, and engaging user experience.

Connected Apps must not

- spam users or the Connected Apps Platform. This includes Connected Apps that send users unsolicited messages or Connected Apps that are repetitive or low-quality.
- copy content from other Connected Apps without adding any original content or value.
- create multiple Connected Apps with highly similar functionality, content and user experience.

> [!NOTE]
> **Note:** Connected Apps that consistently return errors, crash, timeout, or authentication fails when called are prohibited.

## Integration integrity

**Dynamic tool schemas**

Connected Apps must statically declare all callable tool operations in the
console at registration time. Connected Apps must not expose dynamic discovery
functions (e.g., `list_tools`, `get_schema`) or generic runtime dispatchers
(e.g., `execute_command`, `invoke_dynamic_tool`).

**Annotations and human-in-the-loop (HITL)**

Connected app tools that execute sensitive irreversible actions, such as tools
that alter persistent state, modify resources, delete data, alter access
controls, or execute financial transactions, must have `destructiveHint: true`
set under tool annotations or have the operation marked as required confirmation
(HITL) in the console.

**Vertical alignment**

Connected Apps must only provide operations or capabilities that are supported
for the platform vertical they register their connected app in. Connected Apps
that expose unapproved operations for their vertical are prohibited.

**Agent-to-agent integrations**

Connected Apps must not act as a proxy for, connect to, or facilitate
communication with other AI agents that can perform actions autonomously. All
agent-to-agent interactions are strictly prohibited.

## Advertising

Connected app partners must not serve any third party advertising content or
promotional material through their integrations. This does not prohibit the
inclusion of your own brand elements. Connected Apps and their tools must offer
direct, meaningful utility and deliver distinct, standalone value that is
completely free of any advertising content.

## Deceptive behavior

To ensure a safe ecosystem, our standards strictly align with the core
principles of the [Google Unwanted Software
Policy](https://www.google.com/about/unwanted-software-policy.html). Google strictly prohibits all
forms of deception. Your connected app must always maintain honesty and
transparency, and must never mislead users or enable dishonest behavior.

Connected Apps must

- provide an honest, accurate description and, where applicable, a visual representation of their core utility and features across all parts of the connected app and its metadata.
- provide an honest, accurate description of the directory, tools, and permissions; for example, comply with industry naming standards or required naming conventions in relevant specifications.
- define each capability or tool using distinct, unambiguous language.
- operate exactly as described and as a user would reasonably expect.

Connected Apps must not

- enable unlisted functionalities, perform unsolicited actions or actions that are not aligned to the user's intent, cause unexpected side effects, or state features that are missing or incomplete.
- facilitate or provide tools that allow users to deceive others; for example, features designed to generate or spoof official records, identification cards, academic credentials, or financial documents are forbidden.
- contain false or misleading information or claims, including in the description, title, and icon.
- make changes to the user's device settings or features without the user's knowledge and consent.
- modify any device settings or features, if done as a service to third parties or for advertising purposes.
- mislead users into removing or disabling third-party apps or modifying device settings or features.
- promote, create, or perpetuate false or misleading information through deceptively altered imagery, audio, text, or video.
- deceive or influence the model into performing an adverse action or behavior; for example, using overly broad or deceptive capability descriptions to influence the model into selecting the connected app over another connected app.
- reference other tools or create new rules for when or how other tools should be used.
- return system instructions through app responses. All instructions must be available during ingestion time.
- contain hidden, dormant, or undocumented features.
- use evasion techniques, like cloaking or obfuscation to conceal connected app behavior or evade the platform's review process. For example, providing hidden text or content formatted to be illegible to human users while remaining readable by the model.

> [!NOTE]
> **Note:** All principal functions must be disclosed and Connected Apps should not promise a value proposition they do not meet.

## Impersonation and misrepresentation

To protect users and ensure that they can trust a developer's true identity and
the connected app's purpose, Google prohibits developer accounts and Connected
Apps from misrepresenting or concealing their ownership, or primary purpose, or
faking their country of origin to target users. This policy also strictly
forbids coordinating with other Connected Apps, sites or accounts to mislead
users by hiding a connected app's true identity or other material details,
especially when the content involves politics, social issues or matters of
public concern.

Connected Apps and developers must not

- impersonate any person or organization.
- misrepresent or conceal their identity, ownership, primary purpose, or country of origin.
- misrepresent a connected app's functionality or claim capabilities that are technically impossible.
- mislead users about a connected app's relationship to someone else or another connected app.
- engage in coordinated activity with other accounts or sites to mislead users or conceal identity, particularly when connected app content relates to politics, social issues, or matters of public concern.
- wrap or repackage third-party APIs that the developer does not directly own,
  manage, or hold explicit proprietary rights to. The creation of wrapper
  applications, alternative clients, or feature extensions that rely primarily
  on a pass-through of another platform's APIs is strictly prohibited.

  > [!NOTE]
  > **Note:** This does not prohibit applications from utilizing third-party APIs for supporting functions (such as sending email notifications or authenticating users) as part of a larger proprietary service.

## Privacy

We are dedicated to safeguarding user privacy and expect all Connected Apps to
uphold the same high standards. Connected Apps must handle any collected user
data transparently and responsibly.

**User data**

**Data minimization**

Connected Apps must exchange the minimum data necessary to implement the
connected app's current, active features. When defining tool schemas,
parameters, or required inputs, developers must restrict the scope to only what
is strictly required to execute the described functionality of the tool as
submitted to the Platform. Data cannot be used for undisclosed, unimplemented,
or disallowed features or purposes. Connected Apps that rely on permissions must
request the minimum permissions necessary for the connected app to implement its
active features.

Additionally, Connected Apps must never request, collect, or influence the model
to provide user prompts, full conversation context, or conversation history.

**Transparent and accurate notice and control**

You must be transparent about the data you access, collect, use, store, and
share as part of your connected app. You need to maintain a privacy policy that
discloses how the connected app accesses, collects, uses, stores, and shares
user data. Connected Apps must strictly adhere to their published policy at all
times and must never misrepresent what data is being collected or what it is
used for.

Additionally, you must provide user help documentation that explains how users
can manage and delete their data from your app or service.

**Limited use of user data**

Limit your use of user data to only the purposes you have disclosed to the user.
User data may not be exchanged or shared for monetary consideration. You must at
all times comply with privacy and data protection laws applicable in the
jurisdictions in which you offer your products or services.

## Security and abuse

We're committed to providing a safe and secure environment for our users.
Connected Apps that are insecure, or are intended to abuse or misuse any
network, device, or personal data are strictly prohibited.

**Secure data handling**

Developers must implement commercially reasonable administrative, physical, and
technical safeguards to secure user data. This includes using industry-accepted
modern protocols and encryption standards to protect user data both at rest and
in transit.

Connected Apps must

- handle all user data securely, including transmitting it using modern cryptography (for example, over HTTPS).
- ensure keys and key material are managed appropriately, such as stored in a hardware security module or equivalent-strength key management system.
- keep user data and credentials encrypted at rest.

> [!NOTE]
> **Note:** For more information about implementation best practices, see our [AppFunctions](https://developer.android.com/ai/appfunctions) and [MCP server](https://developers.google.com/connected-apps/mcp) developer guides. To register and manage your connected app in [Google Connected Console](https://connected.google.com/console), see the guides for [Android MCP](https://developer.android.com/ai/appfunctions/create-manage-connected-apps) and [MCP servers](https://developers.google.com/connected-apps/gpc).

**Device and network abuse**

We don't allow Connected Apps that interfere with, disrupt, damage or access in
an unauthorised manner the user's device, other devices or computers, servers,
networks, application programming interfaces (APIs) or services, including but
not limited to other apps, Connected Apps, any Google service or an authorised
operator network. Protecting the integrity of the user's device and the broader
ecosystem is paramount.

Connected Apps must not

- obtain unauthorized access or interfere with the user's device, other devices, network, API, or service, other apps or Connected Apps, any Google service, or an authorized operator network.
- introduce or exploit security vulnerabilities.
- access or use of any service or API in a manner that violates its terms of service.
- facilitate the distribution of malicious software.

**Malware and malicious behavior**

Our malware policy is simple: the Connected Apps ecosystem including the
Connected Apps Platform and user devices should be free from malicious
behaviors. Through this fundamental principle we strive to provide a safe
ecosystem for our users and their devices. Malware is any code or instruction
that could put a user, a user's data or a device at risk.

Connected Apps must not

- perform any action that could put a user's data, device or that user at risk.
- display any behavior that is malicious, compromises device integrity, gains unauthorized control, collects or transmits data without adequate disclosure or consent, disseminates spam, or defrauds users.
- enable remote-controlled operations for an attacker to access, use or otherwise exploit an infected device.
- contain hidden code or functionality that performs actions different from the stated purpose or performs undesirable actions against the user.
- manipulate the model into revealing sensitive system instructions or sensitive user data, performing unauthorized actions, or bypassing its own safety protocols.
- escalate privileges or circumvent security, privacy, safety and abuse protections.

## Restricted and prohibited content

To ensure that the Connected Apps Platform experience remains a safe and
respectful platform, we've created standards defining and prohibiting content
that is harmful or inappropriate for our users.

**User generated content**: Any connected app that hosts user generated content
must deploy robust, continuous moderation frameworks. Connected Apps must
moderate their platforms for content that could be harmful to users and to AI
agents. Additionally, Connected Apps must maintain a mandatory Terms of Service
prohibiting objectionable content, and the provision of intuitive, in-app
mechanisms for users to report and block malicious content or abusive actors.

**Child sexual exploitation and abuse**: Connected Apps must not contain any
content that facilitates, depicts, or promotes Child Sexual Abuse Material
(CSAM), grooming, sextortion, child trafficking, or any form of predatory
behavior targeting minors.

**Sexually explicit content**: Connected Apps must not contain pornography,
sexually explicit media, or services intended for sexual gratification. This
includes the prohibition of non-consensual intimate imagery (including
AI-generated deepfakes).

**Hate speech**: Connected Apps must not contain content that promotes violence,
incites hatred, or discriminates against individuals or groups based on
protected characteristics---including race, ethnicity, religion, disability, age,
nationality, veteran status, sexual orientation, gender, or gender identity.

**Violent extremism**: Connected Apps must not contain content related to
violent extremism, or content related to planning, preparing, or glorifying
violence against civilians, such as content that promotes terrorist acts,
incites violence, or celebrates terrorist attacks. We do not permit terrorist
organizations, or other dangerous organizations or movements that have engaged
in, prepared for, or claimed responsibility for acts of violence against
civilians to publish apps on the Connected Apps Platform for any purpose.

**Sensitive events**: We don't allow Connected Apps that capitalize on or are
insensitive toward a sensitive event with significant social, cultural, or
political impact, such as civil emergencies, natural disasters, public health
emergencies, conflicts, deaths, or other tragic events.

**Violence**: Connected Apps must not contain content that depict or facilitate
sensational, shocking, or gratuitous violence.

**Intellectual property**: We don't allow Connected Apps or developer accounts
that infringe on the intellectual property rights of others (including
trademark, copyright, patent, trade secret and other proprietary rights). We
also don't allow Connected Apps that encourage, induce, or facilitate the
unauthorized generation of copyrighted material.

> [!NOTE]
> **Note:** Remember Connected Apps often operate across multiple Google products, platforms, and surfaces. Be sure your connected app remains compliant with applicable surfaces, including but not limited to the [Play Restricted Content
> Policies](https://play.google/developer-content-policy/#!?modal_active=none), [Search Content Policies](https://support.google.com/websearch/answer/10622781) and any other [Service-Specific Policies](https://policies.google.com/terms/service-specific).

## Prohibited goods, services, and functionality

**Illegal activities**: We don't allow Connected Apps that facilitate or promote
illegal activities.

**Age restricted goods \& services**: Connected Apps selling age restricted goods
(e.g., alcohol, tobacco) or that allow age-restricted activities (i.e.
real-money gaming and gambling) must maintain age-gating and verification
procedures.

**Dangerous products**: We prohibit Connected Apps that facilitate the sale of
explosives, firearms, ammunition, or certain firearms accessories.

- Restricted accessories include those that enable a firearm to simulate automatic fire or convert a firearm to automatic fire (for example, bump stocks, gatling triggers, drop-in auto sears, conversion kits), and magazines or belts carrying more than 30 rounds.
- We don't allow apps that provide instructions for the manufacture of explosives, firearms, ammunition, restricted firearm accessories, or other weapons. This includes instructions on how to convert a firearm to automatic, or simulated automatic, firing capabilities.