Skip to main content

Zocks MCP & Claude data privacy

This article explains how Claude handles your data when connected to Zocks via MCP - what's stored, what isn't, and how your Claude plan affects privacy.

Updated over a week ago

What is the Zocks MCP connection?

MCP is an open standard that lets Claude connect to external tools and services. When you connect Zocks via MCP, Claude can read the data in your Zocks account to answer your questions, on demand, and only during your active conversation.

Think of it like giving Claude read-only access to your Zocks workspace. Claude can only read data. It cannot create, edit, or delete anything in Zocks.

How does Claude handle your data?

Your meeting data is retrieved on demand, not downloaded in bulk

When you ask Claude a question, it retrieves the relevant data from Zocks to answer you. Claude doesn't download or index your full Zocks history in the background — it pulls only what's needed for each specific question. Your source data stays in Zocks.

Retrieved data becomes part of your conversation

Here's the important nuance: once Claude retrieves data from Zocks to answer your question, that data becomes part of your Claude conversation. Claude's responses — which often contain sensitive client information like names, financial details, and meeting insights — are saved as part of your conversation history. How that conversation is stored and retained depends on your Claude plan.

Data handling by plan type

Claude Enterprise

  • Anthropic acts as a data processor. Your organization is the data controller and sets the privacy rules.

  • Your admin can configure custom data retention periods (minimum 30 days) in Organization settings → Data and Privacy. By default, data is retained indefinitely unless the admin sets a retention window.

  • Your organization can negotiate a Zero Data Retention (ZDR) agreement with Anthropic, under which Anthropic does not store inputs or outputs beyond what's needed for safety compliance.

  • Your data is not used for model training by default. Admins can also disable the feedback feature organization-wide to prevent any data from being used for training.

Claude Teams

  • Anthropic acts as a data processor. Your organization's privacy configuration applies.

  • Conversations are stored and may be accessible to workspace admins.

  • Your data is not used for model training by default. Team owners can disable the feedback feature to prevent any data from being used for training.

Claude.ai personal plans (Free, Pro, Max)

  • Anthropic acts as the data controller. Anthropic's consumer Privacy Policy applies.

  • Conversations are stored until you delete them. When you delete a conversation, it's removed from your history immediately and deleted from Anthropic's backend within 30 days.

  • By default, Anthropic may use your conversations to improve Claude models. This includes anything Claude retrieved from Zocks during that conversation. You can turn this off: go to Claude.ai → Settings → Privacy and disable "Help improve Claude for everyone."

  • If you opt in to training, de-identified data may be retained for up to 5 years in training pipelines.

If you've discussed sensitive client data in a conversation, delete it when you're done. Don't rely on conversation history as a record of client interactions — use Zocks for that.

Authentication and access control

The Zocks MCP connection uses OAuth authentication — the same standard used by bank-level financial applications. Here's what that means in practice:

  • You authenticate directly with your Zocks account. Claude never sees your Zocks password.

  • Access is scoped to your Zocks account permissions.

  • You can revoke the connection at any time from your Zocks account settings.

Frequently asked questions

Can Claude see all my clients' meeting data at once?

No. Claude retrieves data on demand when you ask a specific question. It doesn't download or index your full Zocks history in the background.

Is my meeting data sent to Anthropic?

When Claude retrieves data from Zocks to answer your question, that data is processed by Anthropic's servers to generate your response. The retrieved data becomes part of your conversation, and how long that conversation is retained depends on your Claude plan (see "Data handling by plan type" above). On Enterprise and Teams plans, your organization controls retention. On personal plans, conversations are stored until you delete them.

Does Anthropic use my data to train AI models?

On Enterprise and Teams plans, no - not by default. On personal plans, Anthropic may use conversations (including data retrieved from Zocks) for model improvement unless you opt out in Settings → Privacy. See the plan-specific details above.

What if I disconnect Zocks MCP? Disconnecting immediately revokes Claude's ability to access your Zocks data. Any conversation already in progress is not affected, but Claude can't retrieve new Zocks data after disconnection.

Can my organization negotiate special data handling terms? Enterprise customers can configure custom retention periods and may negotiate Zero Data Retention (ZDR) agreements with Anthropic. Contact Anthropic's sales team for details.

Questions or concerns?

Contact the Zocks support team at [email protected].

For questions about how Anthropic handles data, visit anthropic.com/privacy , Anthropic's Privacy Center or Anthropic's Trust Center.

Did this answer your question?