Privacy policy

This policy explains the personal information RunTeams processes when providing its website, applications and related services, how information is used and shared, and how you can exercise your rights.

Please understand these points about work content:

The local core and official cloud services handle data separately. Local employee and workflow materials, chats, process records and outputs stay on the executing computer. The local core can run independently. When you use official accounts, the marketplace, organizations or transactions, related platform records are handled by cloud services. Deliberate sharing and feedback send the relevant content to the selected service.

Agents run on your computer and also connect to external services. They may read authorized files, connected services and task materials, and send necessary content to the agent runtime, model service or other providers you choose. Local execution does not mean all information stays on your computer.

Team sharing and public publishing have different audiences. Personal spaces, organizations and the public marketplace are subject to their permissions and publication settings. Check for private information before publishing.

This policy describes information handling. Where separate consent is legally required, browsing or continuing to use the service does not replace it.

Information comes from materials you provide, your use of the service, services you authorize, and your organization. The categories depend on the features you use.

Account and profile information: sign-in email, account identifiers, login status, and any name or avatar you provide, used for authentication, account management and service notices. Google sign-in also involves authentication information. Google sign-in and Gmail authorization are separate actions.

Work content and results: messages, prompts, task inputs, employee and workflow configurations, skill materials, attachments, work records and deliverables, used to save, execute, collaborate on and deliver your work. This may include your or other people’s personal information.

Team and permission information: organization names, members, invitations, roles, workspace resources, task initiators and executing computers, used for collaboration, access and traceability.

Device and execution information: account-linked device identifiers, names, app and environment information, and execution, error and security records. These help connect devices, recover work and investigate problems. Logs may include tool inputs and outputs, email, file contents, paths and errors; they are not necessarily anonymous. Diagnostics and formal work evidence may have different retention rules.

Connection and authorization information: services, account identifiers, access scopes, connection status, associations, action requests and receipts. Requests may contain message bodies, recipients or attachments. Local access tokens and agent credentials are managed separately from cloud records.

Update and resource requests: update checks and downloads may disclose network addresses, versions and other request details to the recipient. Plugin avatars and website icons may send public plugin ownership, domains and network information to GitHub or Google.

Transactions and entitlements: orders, plans, validity periods, payment and refund status, payment-provider customer identifiers, wallets and marketplace transactions, used to confirm access, reconcile accounts and resolve disputes. Payment information is handled by the provider identified at checkout.

Support: contact details, messages, screenshots and materials you provide, plus network, device and session information processed by support tools. Do not send passwords, keys or unrelated sensitive information.

Public content: materials, asset descriptions and publisher information you choose to publish. Other people may view, save or redistribute public content outside the service.

Features may be unavailable without information necessary for authentication, permissions or execution. Optional profile information, materials and services remain your choice. Not connecting Gmail does not prevent work that does not need email.

Only provide materials you are entitled to process. For other people’s information, medical information, identity documents or other specially protected information, check the legal basis, purpose and protection requirements first.

RunTeams cloud services and your computer

The local core stores employees, workflows, working directories, chats, process records and outputs on the executing computer and can be used independently. When you choose official accounts, the marketplace, organizations or transactions, related account, permission, publication, transaction and asset records use cloud services and access controls. Deliberate sharing, publication and feedback send the relevant content for that feature. Refer to Supabase’s privacy policy and the cross-border processing section.

Local work content is persistent data, not a cache that only becomes valid after upload. Display preferences, credentials, native agent configuration, file grants and recovery records also reside on the device. Backups and deliberate sharing are separate; sharing transfers selected content. Clearing caches or uninstalling does not automatically delete cloud records.

In the official desktop app, RunTeams refresh tokens use system credential storage; short-lived access tokens remain in the running process. Tool credentials managed by RunTeams are stored in local files protected by filesystem permissions. Selected agents or extensions manage their own accounts and authorizations. These credentials do not all use the system keychain or identical encryption.

Agent runtimes and external services

Supported integrations include OpenAI Codex and Anthropic Claude Code. They and their connected services may process task prompts, file excerpts, conversation context, tool calls and responses. Scope depends on the task, environment, extensions, permissions and service settings.

Automatic verification may send the task, output, deliverables and relevant evidence to the selected runtime to check and continue the work. This is task processing, not the same as training a general-purpose model.

Third-party services you subscribe to also have their own terms, policies, age requirements and data settings. Disconnecting RunTeams does not close those accounts, cancel subscriptions or delete information already received.

Information obtained through Google Workspace APIs is subject to Google Workspace API user-data and developer policies and the Google API Services User Data Policy, including Limited Use. It must not be used for advertising targeting or training or improving general-purpose AI models. Human access and third-party disclosure are limited to permitted purposes and circumstances.

Support and optional mobile features

Feedback in the official version is handled by the RunTeams message service and linked to the submitting account. Messages, images and other materials you submit are used to address inquiries and reply. Submitting a rating for a chat reply also includes the conversation through that reply, attachment references, relevant images and some runtime information to review the reply. These materials may contain work content, local paths or personal information.

Mobile features are being revised. The confirmed direction is limited encrypted or redacted progress information; fields, key mechanisms and remote operations are not yet settled. This does not establish full chat or file sync or remote takeover. Actual processing and choices must be explained before release.

Website sign-in

Website sign-in tokens are stored in the current tab's session storage, or only in page memory when session storage is unavailable. The website does not read desktop sign-in tokens. Restoring a session still checks the user's identity with the authentication service.

Usage analytics

When analytics is enabled, we use PostHog to receive controlled events about page visits, sources, download clicks, sign-in stages, and basic performance and errors. Events use random browser-generated identifiers without attaching RunTeams account IDs, email addresses, work content or full URLs containing query parameters. Analytics requests still transmit network information, and the provider may infer a location from an IP address. The website respects browser Do Not Track and Global Privacy Control signals. See the PostHog Privacy Policy for details.

We use information to provide requested services: authenticate accounts, save content, run tasks, deliver results, support collaboration and connect external services. Other purposes include transactions, support, abuse prevention, security and legal obligations.

Third-party data use depends on your account, service and settings. Review your agent provider’s information and settings before assigning work. RunTeams authorization does not replace third-party terms or privacy choices. Gmail data remains subject to the Limited Use requirements above.

Where applicable law requires it, processing relies on contractual necessity, legal obligations or valid consent. Separate consent is obtained when required. Withdrawal does not affect lawful processing based on consent before withdrawal.

When an enterprise customer determines the purposes of processing work content, applicable data-processing arrangements and lawful instructions apply. Accepting the terms is not blanket consent for every personal-information activity.

Information may be provided as needed for the service or required by law to:

Authorized team members, according to organization roles and resource permissions. Membership does not expose all personal content or device credentials.

Infrastructure, authentication, payment and support providers, including Supabase and PostHog, and the payment processor identified at checkout, such as Stripe. Roles depend on actual services and contracts.

External services you select or authorize, including OpenAI Codex, Anthropic Claude Code, Gmail and other extensions or destinations. Outbound content depends on actual permissions and execution settings.

Optional feature and resource services, such as Apple APNs for mobile push, GitHub and Google for icons, and update or download services. They receive the requests needed when those paths are used.

Visitors to public areas, for information you choose or authorize someone to publish.

Recipients legally entitled to information. We check applicable requirements and provide notice where permitted.

Display preferences such as themes and window layouts use local storage; some other business preferences use cloud storage. Desktop login tokens are managed by the local account service rather than directly exposed to the front end.

Website language preferences may be stored in local storage and a cookie, while theme preferences use local storage. Website sign-in uses the current tab's session storage. When usage analytics is enabled, browser-generated random identifiers and visit-session information are also used.

The in-app feedback interface may keep unsent text, images and unconfirmed requests in session storage to restore the original message state. Processing after feedback is sent is described in the data-flow section.

You can manage cookies and local storage in browser or device settings. Restricting necessary storage may affect sign-in, preferences or features. Nonessential technologies requiring consent under applicable law should only be enabled after consent.

We retain information for the purposes and periods needed to provide the service, taking account of legal obligations, outstanding transactions, disputes and security.

Work content

Saved tasks, chats, files and results are retained as needed for the service. Trash normally retains items for 30 days, with cleanup triggered during service processing. Document protection, running work or other dependencies may affect timing; the trash period does not mean every copy is deleted on the same day.

Execution and transaction records

Execution, error, security and support records are retained for their purposes. Transaction and accounting records also follow applicable financial, tax and dispute requirements. Formal work evidence and temporary diagnostics are managed separately.

Deletion and stopping use

Canceling a subscription, leaving or closing an organization, disconnecting a service, deleting local files or uninstalling does not automatically delete your account or all cloud records. Closing an organization revokes member access while retaining business data. Organization records and personal information may follow different rules.

Use available deletion features or contact support with a personal-information request. We verify identity and applicable requirements. Information that must legally be retained is kept only for the relevant purposes and obligations.

Your location, the service’s operating location, cloud deployment and third-party processing locations may differ. Cloud storage, agent runtimes and connected services may process information outside your region.

Review the categories, recipients and provider policies listed here. Where applicable law requires it, necessary notices, contractual arrangements, assessments or consent will be completed before the relevant transfer.

Available features let you edit your profile, change team permissions, stop automations, disconnect services and choose whether to publish. Stopping future access is different from deleting retained content.

Depending on applicable law and circumstances, rights may include access or a copy, correction, deletion, restriction, objection, portability, withdrawal of consent and complaints to a supervisory authority. Some regions also provide rights to opt out of sale, certain sharing or targeted advertising, and rights of appeal.

To exercise rights, contact support with your request and account. We may seek proportionate information to verify identity or authority. Response periods, lawful extensions, fee exceptions and refusal reasons follow applicable law.

For work content whose purposes are determined by an enterprise customer, contact that organization first. We assist under applicable processing arrangements. Organization membership alone does not grant access to local credentials you have not shared.

You must meet local legal requirements and the age requirements of chosen agents and services. Complete parental consent, assistance or verification where legally required. Parental permission cannot bypass a third-party service’s minimum age or account rules.

Guardians may contact support to request access, withdrawal or deletion. We verify identity, guardianship and applicable requirements and respond according to law.

If collection does not meet applicable child-protection requirements, we take appropriate steps to stop processing and delete or restrict retention as required.

The product checks account and workspace permissions. Remote business requests use HTTPS with server certificate verification. Local login credentials have separate protections; the mobile progress mechanism is still to be determined. Authorized cloud services can process ordinary business content. This is not full end-to-end encrypted storage where only you hold decryption keys.

Protect your accounts, devices, file permissions and agent configuration. For personal-information security incidents, we meet applicable investigation, response and notification obligations.

The update date appears on this page. Material changes to purposes, recipients or rights will be communicated appropriately in advance or within the period required by law. Where renewed consent is required, it will be obtained before processing.

For questions about this policy or personal-information handling, email support@runteams.ai. Identify the account, issue and help needed without sending passwords, keys or unrelated sensitive information.

Last updated: November 3, 2026.