mirror of
https://github.com/zed-industries/zed.git
synced 2026-07-09 16:00:35 +00:00
Document provider safety retention for designated Zed-hosted models (#58967)
Anthropic now requires limited data retention for models it designates as covered models (Mythos-class models, such as Claude Fable 5): prompts and outputs are retained for 30 days for trust and safety purposes, on every platform where these models are offered, including platforms with zero-data-retention agreements. See [Anthropic's public policy](https://support.claude.com/en/articles/15425996-data-retention-practices-for-mythos-class-models). Our AI privacy docs previously described zero data retention as applying to all Zed-hosted models unconditionally. This PR updates them to document the exception: - `docs/src/ai/privacy-and-security.md`: Adds a "Provider Safety Retention for Designated Models" section covering what is retained, by whom, for how long, that no-training commitments still apply, and that bringing your own key does not avoid retention for covered models. Qualifies the intro, request-path table, and provider commitments table accordingly. - `docs/src/ai/ai-improvement.md`: Qualifies the zero-data-retention statements and links to the new section. - `docs/src/business/privacy.md`: Qualifies the zero-data-retention statement and links to the new section. Release Notes: - N/A
This commit is contained in:
morgankrey
GitHub
parent
f1ad1e7fd5
commit
3d2f7adb95
5 changed files with 68 additions and 33 deletions
|
|
@ -6,10 +6,13 @@ description: Understand opt-in AI feedback ratings, Edit Prediction training dat
|
|||
# Feedback and Training Data
|
||||
|
||||
Normal AI requests are not retained by Zed. For
|
||||
[Zed-hosted models](../account/zed-hosted-models.md), provider agreements require
|
||||
zero data retention and prohibit training on your prompts or code context. This
|
||||
page covers the cases where Zed may retain AI data because you explicitly shared
|
||||
it or opted in.
|
||||
[Zed-hosted models](../account/zed-hosted-models.md), provider agreements
|
||||
prohibit training on your prompts or code context and require zero data
|
||||
retention, except for
|
||||
[provider-designated models with safety retention](./privacy-and-security.md#provider-safety-retention),
|
||||
such as Anthropic's Mythos-class models.
|
||||
This page covers the cases where Zed may retain AI data because you explicitly
|
||||
shared it or opted in.
|
||||
|
||||
AI features in Zed include:
|
||||
|
||||
|
|
@ -23,7 +26,9 @@ For the broader request path and provider data boundaries, see
|
|||
|
||||
## Zed-Hosted Model Commitments {#data-retention-and-training}
|
||||
|
||||
Zed-hosted model zero-data-retention and no-training commitments are documented
|
||||
Zed-hosted model zero-data-retention and no-training commitments, including the
|
||||
exception for provider-designated models with safety retention (such as
|
||||
Anthropic's Mythos-class models), are documented
|
||||
on [AI Privacy](./privacy-and-security.md#data-retention-and-training).
|
||||
|
||||
## Response Ratings and Feedback {#ai-feedback-with-ratings}
|
||||
|
|
|
|||
|
|
@ -11,39 +11,69 @@ including [Zed Agent](./zed-agent.md), [Edit Prediction](./edit-prediction.md),
|
|||
[Git commit generation](../git.md#ai-support-in-git).
|
||||
|
||||
Zed does not retain your prompts or code context by default. For
|
||||
[Zed-hosted models](../account/zed-hosted-models.md), Zed has zero-data-retention
|
||||
and no-training commitments from model providers for inference requests. Zed only
|
||||
retains AI data when you explicitly share feedback or opt in to training data
|
||||
collection.
|
||||
[Zed-hosted models](../account/zed-hosted-models.md), Zed has no-training
|
||||
commitments from model providers, and provider agreements require zero data
|
||||
retention for inference requests except for
|
||||
[provider-designated models with safety retention](#provider-safety-retention),
|
||||
such as Anthropic's Mythos-class models.
|
||||
Zed only retains AI data when you explicitly share feedback or opt in to
|
||||
training data collection.
|
||||
|
||||
## AI Request Paths {#ai-request-paths}
|
||||
|
||||
| Path | Who handles model requests | What to know | Details |
|
||||
| ------------------------------------------------------------ | ------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- |
|
||||
| [Zed-hosted models](../account/zed-hosted-models.md) | Zed routes requests to hosted model providers | Provider agreements require zero data retention and prohibit training on your prompts or code context for inference requests. | [Zed-hosted model commitments](#data-retention-and-training) |
|
||||
| [Provider API keys](./use-api-access.md) | The configured provider | The provider handles requests under its own terms. Provider keys saved through Zed are stored in the system keychain, not in `settings.json`. | [Use API Access](./use-api-access.md) |
|
||||
| [Existing subscriptions](./use-an-existing-subscription.md) | The subscription provider | The provider handles requests under the subscription terms. | [Use an Existing Subscription](./use-an-existing-subscription.md) |
|
||||
| [Gateways](./use-a-gateway.md) | The configured gateway and upstream providers | The gateway and upstream providers handle requests under their own terms. | [Use a Gateway](./use-a-gateway.md) |
|
||||
| [Local models](./use-a-local-model.md) | The local server or self-hosted endpoint | The local server handles requests according to how you configured that server. | [Use a Local Model](./use-a-local-model.md) |
|
||||
| [External Agents](./external-agents.md) | The External Agent and its configured providers | The External Agent handles model requests under its own terms. Tool and MCP behavior depends on agent and ACP configuration. | [External Agents](./external-agents.md) |
|
||||
| [Terminal Threads](./terminal-threads.md) | The CLI or TUI running in the terminal | The CLI or TUI owns its auth, model routing, tools, instructions, MCP configuration, and data handling. | [Terminal Threads](./terminal-threads.md) |
|
||||
| [Edit Prediction](./edit-prediction.md) | The selected edit prediction provider | Each keystroke can send local editing context to the selected provider. Zeta requests are processed transiently unless training data collection is enabled; third-party providers follow their own terms. | [Edit Prediction](./edit-prediction.md), [Feedback and Training Data](./ai-improvement.md) |
|
||||
| [Agent tools](./tools.md), [MCP](./mcp.md), and integrations | Zed, configured MCP servers, and external systems | Tools can read, edit, search, run commands, fetch URLs, or call external systems depending on profile, MCP server, and tool permission settings. | [Agent Profiles](./agent-profiles.md), [Tool Permissions](./tool-permissions.md), [MCP](./mcp.md) |
|
||||
| Project trust and instructions | Zed and the trusted worktree | Project-local instructions and skills are loaded from trusted worktrees. External Agents and Terminal Threads may read their own instruction files. | [Worktree Trust](../worktree-trust.md), [Skills](./skills.md), [Instructions](./instructions.md) |
|
||||
| Path | Who handles model requests | What to know | Details |
|
||||
| ------------------------------------------------------------ | ------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- |
|
||||
| [Zed-hosted models](../account/zed-hosted-models.md) | Zed routes requests to hosted model providers | Provider agreements prohibit training on your prompts or code context and require zero data retention for inference requests, except for provider-designated models with safety retention, such as Anthropic's Mythos-class models. | [Zed-hosted model commitments](#data-retention-and-training) |
|
||||
| [Provider API keys](./use-api-access.md) | The configured provider | The provider handles requests under its own terms. Provider keys saved through Zed are stored in the system keychain, not in `settings.json`. | [Use API Access](./use-api-access.md) |
|
||||
| [Existing subscriptions](./use-an-existing-subscription.md) | The subscription provider | The provider handles requests under the subscription terms. | [Use an Existing Subscription](./use-an-existing-subscription.md) |
|
||||
| [Gateways](./use-a-gateway.md) | The configured gateway and upstream providers | The gateway and upstream providers handle requests under their own terms. | [Use a Gateway](./use-a-gateway.md) |
|
||||
| [Local models](./use-a-local-model.md) | The local server or self-hosted endpoint | The local server handles requests according to how you configured that server. | [Use a Local Model](./use-a-local-model.md) |
|
||||
| [External Agents](./external-agents.md) | The External Agent and its configured providers | The External Agent handles model requests under its own terms. Tool and MCP behavior depends on agent and ACP configuration. | [External Agents](./external-agents.md) |
|
||||
| [Terminal Threads](./terminal-threads.md) | The CLI or TUI running in the terminal | The CLI or TUI owns its auth, model routing, tools, instructions, MCP configuration, and data handling. | [Terminal Threads](./terminal-threads.md) |
|
||||
| [Edit Prediction](./edit-prediction.md) | The selected edit prediction provider | Each keystroke can send local editing context to the selected provider. Zeta requests are processed transiently unless training data collection is enabled; third-party providers follow their own terms. | [Edit Prediction](./edit-prediction.md), [Feedback and Training Data](./ai-improvement.md) |
|
||||
| [Agent tools](./tools.md), [MCP](./mcp.md), and integrations | Zed, configured MCP servers, and external systems | Tools can read, edit, search, run commands, fetch URLs, or call external systems depending on profile, MCP server, and tool permission settings. | [Agent Profiles](./agent-profiles.md), [Tool Permissions](./tool-permissions.md), [MCP](./mcp.md) |
|
||||
| Project trust and instructions | Zed and the trusted worktree | Project-local instructions and skills are loaded from trusted worktrees. External Agents and Terminal Threads may read their own instruction files. | [Worktree Trust](../worktree-trust.md), [Skills](./skills.md), [Instructions](./instructions.md) |
|
||||
|
||||
## Zed-Hosted Model Commitments {#data-retention-and-training}
|
||||
|
||||
For Zed-hosted models, Zed has commitments from model providers that prohibit
|
||||
training on your prompts or code context and require zero data retention for
|
||||
inference requests. The public provider documents linked below describe provider
|
||||
programs or default API terms; Zed-hosted model requests are governed by Zed's
|
||||
provider agreements.
|
||||
inference requests, except for
|
||||
[provider-designated models with safety retention](#provider-safety-retention),
|
||||
such as Anthropic's Mythos-class models. The public provider documents linked below describe provider programs or default
|
||||
API terms; Zed-hosted model requests are governed by Zed's provider agreements.
|
||||
|
||||
| Provider | No training reference | Zero-data-retention reference |
|
||||
| --------- | ------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| Anthropic | [Yes](https://www.anthropic.com/legal/commercial-terms) | [Yes](https://privacy.anthropic.com/en/articles/8956058-i-have-a-zero-data-retention-agreement-with-anthropic-what-products-does-it-apply-to) |
|
||||
| Google | [Yes](https://cloud.google.com/terms/service-terms) | [Yes](https://cloud.google.com/terms/service-terms), see Service Terms sections 18 and 20(h) |
|
||||
| OpenAI | [Yes](https://openai.com/enterprise-privacy/) | [Yes](https://platform.openai.com/docs/guides/your-data) |
|
||||
| Provider | No training reference | Zero-data-retention reference |
|
||||
| --------- | ------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| Anthropic | [Yes](https://www.anthropic.com/legal/commercial-terms) | [Yes](https://privacy.anthropic.com/en/articles/8956058-i-have-a-zero-data-retention-agreement-with-anthropic-what-products-does-it-apply-to), except [covered models](#provider-safety-retention) |
|
||||
| Google | [Yes](https://cloud.google.com/terms/service-terms) | [Yes](https://cloud.google.com/terms/service-terms), see Service Terms sections 18 and 20(h) |
|
||||
| OpenAI | [Yes](https://openai.com/enterprise-privacy/) | [Yes](https://platform.openai.com/docs/guides/your-data) |
|
||||
|
||||
### Provider Safety Retention for Designated Models {#provider-safety-retention}
|
||||
|
||||
Some providers require limited data retention for specific models as a condition
|
||||
of offering them, on every platform where those models are available. Anthropic
|
||||
retains prompts and outputs for models it designates as covered models (its
|
||||
Mythos-class models, such as Claude Fable 5) for 30 days for trust and safety
|
||||
purposes. Zed cannot opt out of this retention; it applies wherever these models
|
||||
are served. See
|
||||
[Anthropic's data retention practices for Mythos-class models](https://support.claude.com/en/articles/15425996-data-retention-practices-for-mythos-class-models).
|
||||
|
||||
For these models:
|
||||
|
||||
- The no-training commitment still applies. Retained data is used for safety
|
||||
review, not model training.
|
||||
- Zed does not retain your prompts or outputs. Retention happens at the
|
||||
provider, under the provider's documented access controls and deletion
|
||||
timelines.
|
||||
- All other Zed-hosted models keep zero-data-retention handling.
|
||||
|
||||
If you don't want provider-side retention, use a model that the provider has
|
||||
not designated for safety retention. Switching to
|
||||
[your own API key](./use-api-access.md) or
|
||||
[subscription](./use-an-existing-subscription.md) does not avoid this retention
|
||||
for covered models, because providers apply it on every platform where those
|
||||
models are offered.
|
||||
|
||||
## AI Data Retained by Zed {#ai-data-retained-by-zed}
|
||||
|
||||
|
|
|
|||
|
|
@ -36,7 +36,7 @@ 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 zero-data-retention and no-training commitments with these providers. See [AI Privacy](../ai/privacy-and-security.md#data-retention-and-training) for details.
|
||||
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](../ai/privacy-and-security.md#provider-safety-retention), such as Anthropic's Mythos-class models. See [AI Privacy](../ai/privacy-and-security.md#data-retention-and-training) for details.
|
||||
|
||||
[Bring-your-own-key](../ai/llm-providers.md), [gateways](../ai/use-a-gateway.md), [local or self-hosted models](../ai/use-a-local-model.md), [External Agents](../ai/external-agents.md), and [Terminal Threads](../ai/terminal-threads.md) are subject to each provider, gateway, server, agent, or CLI's own terms.
|
||||
|
||||
|
|
|
|||
|
|
@ -3,11 +3,11 @@ title: Privacy Policy
|
|||
slug: privacy-policy
|
||||
---
|
||||
|
||||
**Last Updated**: March 2, 2026
|
||||
**Last Updated**: June 9, 2026
|
||||
|
||||
## Summary
|
||||
|
||||
Zed collects user information (name, email, username, IP address, etc.), device and usage data, and web analytics to improve our products and services, as well as payment details if you subscribe. We do not store your source code and we process but do not store audio/video/screenshare from collaboration sessions.
|
||||
Zed collects user information (name, email, username, IP address, etc.), device and usage data, and web analytics to improve our products and services, as well as payment details if you subscribe. We do not store your source code and we process but do not store audio/video/screenshare from collaboration sessions. For a limited set of provider-designated AI models (such as Anthropic's Mythos-class models), the model provider retains prompts and outputs for a limited period for safety purposes; see our [AI Privacy documentation](https://zed.dev/docs/ai/privacy-and-security).
|
||||
|
||||
- **We don't sell your data and we don't allow AI providers to train on it**
|
||||
- **You control Zed editor telemetry** — disable it in Settings or learn more at [https://zed.dev/docs/telemetry](/telemetry)
|
||||
|
|
|
|||
|
|
@ -94,7 +94,7 @@ _These subprocessors apply only when customers opt to use Zed's hosted AI models
|
|||
|
||||
| Subprocessor | Purpose | Data Location |
|
||||
| ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------- |
|
||||
| **Anthropic** | Requests may be sent to Anthropic even if you have another provider's model selected in chat (e.g. for summarization or generating git commit messages). We have a zero data retention agreement with Anthropic. | United States |
|
||||
| **Anthropic** | Requests may be sent to Anthropic even if you have another provider's model selected in chat (e.g. for summarization or generating git commit messages). We have a zero data retention agreement with Anthropic, except for models Anthropic designates as covered models, for which Anthropic retains prompts and outputs for 30 days for trust and safety purposes ([Anthropic's policy](https://support.claude.com/en/articles/15425996-data-retention-practices-for-mythos-class-models)). | United States |
|
||||
| **Baseten** | Inference infrastructure for Edit Predictions | United States |
|
||||
| **Exa Labs** | AI-powered contextual search and retrieval | United States |
|
||||
| **Google (Vertex)** | Requests may be sent to Google even if you have another provider's model selected in chat (e.g. for summarization). We have a zero data retention agreement with Google. | United States |
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue