AI & Integrations Terms

Version 2.0 · Effective October 2, 2026 · Last updated October 2, 2026

These terms supplement the Terms of Service. They govern GrowingWhale's AI Features, our MCP server, automations and third-party connectors. If they conflict with the Terms of Service, these terms control for those features. Capitalized terms have the meanings in the Terms of Service.

Part A: AI features

A1. What they are

GrowingWhale uses third-party AI models, including Anthropic's Claude and other providers we may use from time to time, for content and ad copy, review replies, chat assistants, page assembly, insights and agentic actions. AI Features depend on our providers and may be unavailable in some countries or change when providers change their models.

A2. You are the final reviewer

• AI output can be wrong, outdated, biased, offensive or infringing, and it can invent facts.
• You must review and approve AI-generated content before publishing, sending or relying on it.
• If you turn on automatic posting of AI-drafted replies (for example, to Google or Meta reviews), you accept responsibility for every reply posted and for compliance with those platforms' policies.
• You are solely responsible for what you publish, including claims about your products, prices, health, finance and legal matters.
• AI Features don't provide legal, financial, medical, tax or other professional advice.

A3. Ownership of outputs

As between you and us, you own the outputs generated for you, to the extent the law allows. Similar outputs may be generated for others. We don't guarantee outputs are unique, protected by copyright, or free of third-party rights; check them for others' trademarks, copyrighted material and people's names or likenesses before use.

A4. No training on your data

We don't use your Member Data or prompts to train generalized AI models. We use AI providers under terms that prohibit them from training on your inputs and outputs. This doesn't cover AI services you connect with your own API keys, including in your own workflows; their terms apply.

A5. Prohibited AI use

You must not use AI Features to:

• generate unlawful, deceptive, defamatory or harassing content, or fake reviews or testimonials;
• impersonate people, or create realistic images, audio or video of real people without their consent;
• make automated decisions with legal or similarly significant effects on individuals, such as about credit, housing, employment or insurance;
• violate the usage policies of the model provider, including Anthropic's Usage Policy.

A6. AI chat, voice and disclosure to your End Users

• If you put an AI chat assistant on your website or in your messages, you must tell End Users they are interacting with AI, as several US states and the EU AI Act require. Keep our default "AI assistant" label turned on unless you disclose it another way.
• Never present AI as a human or deny it is AI when asked.
• AI voice agents. If you have an AI agent call your customers or answer your calls, it acts for you and you are the caller. You must follow Section 8.8 of the Terms of Service, including consent, do-not-call checks, an AI disclosure at the start of each call, and recording consent.
• Give an AI agent clear instructions and limits on what it may say, promise, book or do. You are responsible for its statements and actions, including bookings, quotes, refunds and commitments it makes.
• We may limit an AI agent's call volume, hours or destinations, and end calls or tasks that appear unlawful, abusive or outside your instructions.
• Label AI-generated images, audio or video of people as AI-generated where the law requires.

Part B: Our MCP server (mcp.growingwhale.ai)

B1. What it does

Our Model Context Protocol (MCP) server lets you connect Claude and other MCP-compatible AI clients to your GrowingWhale workspace. The AI client can then read data and take actions, such as creating pages, posts or CRM records, on your behalf. Claude is a trademark of Anthropic, PBC. GrowingWhale is not affiliated with or endorsed by Anthropic.

B2. Authorization

• Access requires OAuth sign-in and your explicit consent to the listed scopes.
• Tokens are scoped to your workspace and can be revoked anytime in Settings → Connected Apps or in your AI client.

B3. You are responsible for AI-initiated actions

• Actions an AI client takes through your MCP connection are treated as actions by you.
• We may require you to confirm high-impact actions before they run, such as publishing content, messaging many recipients, changing ad budgets or deleting records. Keep confirmations turned on.
• Don't grant MCP access to AI clients or people you don't trust.
• We are not liable for actions an AI client takes through your authorized connection, including deletions, publications or messages sent, except where the loss is caused by our gross negligence or willful misconduct.

B4. Third-party AI clients

Claude apps and other AI clients are provided by their own companies, which store conversations under their own terms. We don't control them and aren't responsible for their availability, security or data practices.

B5. Security and our controls

We apply rate limits, input validation and least-privilege tool design, and we monitor the MCP server for abuse. We may block specific tools or actions, revoke tokens, or suspend MCP access to protect you, other Members or the Service. Prompt injection and similar attacks on AI systems are an evolving risk, so:

• avoid feeding untrusted content to agents with write access;
• keep a person in the loop for high-impact actions.

B6. Data and retention

We log the AI client's name, granted scopes, tool calls and timestamps for up to 90 days to run and secure the MCP server. Data the MCP server returns to your AI client at your request is then held by that client's provider under its terms.

Part C: Automations (n8n workflows)

C1. How they run

Workflows may run on n8n infrastructure operated by us or our hosting providers, or on your own n8n instance if you connect one.

C2. Your responsibility

• You design, enable or approve each workflow, including its triggers, recipients, data mappings and frequency.
• You are responsible for making sure automations comply with law and platform terms, including the messaging rules in Section 8 of the Terms of Service.
• You are responsible for their results, including mistaken sends, duplicate records, data overwrites and API charges.
• Test workflows before turning them on for live data.

C3. No guarantee of execution

Automations depend on third-party APIs, webhooks, network conditions and quotas. We don't guarantee that any workflow will run on time, exactly once or at all, and we aren't liable for missed, delayed, duplicated or failed runs, or for data loss caused by third-party failures.

C4. Self-hosted n8n

If you connect your own n8n instance, you are solely responsible for its hosting, security, updates, licensing, credentials and data.

C5. Limits

We may throttle, pause or disable workflows that exceed plan limits, create security risk, break platform or carrier rules, or harm the Service's performance.

Part D: Third-party connectors

D1. Connecting apps

You can connect third-party services, such as Google, Meta, Microsoft, Stripe, Zoho, HubSpot, Shopify, Slack, webhooks and custom APIs, per Member account. Each connection is authorized by the person who owns that third-party account.

D2. Your authorization and representations

When you connect a service, you confirm that you own it or are authorized to connect it, and that sharing data between it and GrowingWhale is permitted by its terms and by law.

D3. What we do and don't control

• We access connected services only within the scopes you grant, and only to perform features you use.
• We don't control third-party services. They may change APIs, prices, permissions or policies, suspend your account, or suffer outages and breaches. We aren't responsible for these events or their consequences, except as Section 10.3 of the Terms of Service provides.
• Features that depend on a third-party approval, such as Meta App Review or Google verification, may be limited or unavailable at any time.

D4. Credentials

• We use safeguards designed to protect API keys, OAuth tokens and webhook secrets, including encryption, and keep them separate for each Member.
• Rotate keys if you suspect compromise and tell us at security@growingwhale.com.
• Don't paste credentials into free-text fields, prompts or content areas.

D5. Data flowing out

Once data is sent to a third-party service at your direction, that service's terms govern it, and we aren't responsible for how it handles it.

Part E: Liability

The disclaimers, liability limits and indemnity in Sections 17 to 19 of the Terms of Service apply in full to AI Features, the MCP server, automations and connectors.

Contact

Shopinzip Inc · 222 Highview Drive, Clifton, NJ 07013, USA · legal@growingwhale.com