# Privacy for Business - Zed Business

> How Zed Business handles data privacy across your organization, including enforced protections for prompts and training data.

- 网址：https://funcoding.ai/agents/zed/business/privacy/
- 来源：Zed 官方文档原文（英文），Apache-2.0 / GPL-3.0 许可，同步于 2026-10-11
- 官方原文：https://zed.dev/docs/business/privacy

---
Zed Business removes the per-member data-sharing options that Free and Pro
expose. These protections are on by default for every Business organization.
Administrators can adjust them from
[Admin Controls](https://funcoding.ai/agents/zed/business/admin-controls/); individual members can't opt in or out.

## What's enforced by default

For all members of a Zed Business organization:

- **No prompt sharing:** Conversations and prompts are never shared with Zed.
  Members can't opt into
  [AI feedback via ratings](https://funcoding.ai/agents/zed/ai/ai-improvement/#ai-feedback-with-ratings).
  Administrators can enable Agent Thread Feedback to allow this.
- **No training data sharing:** Code context is never shared with Zed for
  [Edit Prediction model training](https://funcoding.ai/agents/zed/ai/ai-improvement/#edit-predictions).
  Members can't opt in individually. Administrators can enable Edit Prediction
  Feedback to allow this.

These protections are enforced server-side and apply to all org members.

## How individual plans differ

On Free and Pro, data sharing is opt-in:

- Members can rate AI responses, which shares that conversation with Zed.
- Members can opt into Edit Prediction training data collection for open source projects.

Neither option is available to Zed Business members.

## What data still leaves the organization

These controls cover what Zed stores and trains on. They don't change how AI inference works: when members use Zed's hosted models, prompts and code context are still sent to the relevant provider (Anthropic, OpenAI, Google, etc.) to generate responses. Zed maintains no-training commitments with these providers, and zero-data-retention commitments for all models except [provider-designated models with safety retention](https://funcoding.ai/agents/zed/ai/privacy-and-security/#provider-safety-retention), such as Anthropic's Covered Models. See [AI Privacy](https://funcoding.ai/agents/zed/ai/privacy-and-security/#data-retention-and-training) for details.

[Bring-your-own-key](https://funcoding.ai/agents/zed/ai/llm-providers/), [gateways](https://funcoding.ai/agents/zed/ai/use-a-gateway/), [local or self-hosted models](https://funcoding.ai/agents/zed/ai/use-a-local-model/), [External Agents](https://funcoding.ai/agents/zed/ai/external-agents/), and [Terminal Threads](https://funcoding.ai/agents/zed/ai/terminal-threads/) are subject to each provider, gateway, server, agent, or CLI's own terms.

## Additional admin controls

Administrators have additional options in [Admin Controls](https://funcoding.ai/agents/zed/business/admin-controls/):

- Disable Zed-hosted models entirely via the Zed Model Provider toggle, so no
  prompts reach Zed's infrastructure
- Disable Edit Predictions org-wide
- Disable Edit Prediction Feedback
- Disable Agent Thread Feedback
- Disable real-time collaboration

See [Admin Controls](https://funcoding.ai/agents/zed/business/admin-controls/) for the full list.