Zero Data Retention
A data compliance arrangement between enterprises and model vendors in which the vendor undertakes not to store any data sent by the customer and not to use it for training. It is an important consideration in selecting enterprise-grade AI applications.
Core Facts
Zero retention does not equal zero learning—the fact that data is not stored does not mean that patterns and information extracted from it will not influence the model in some form
AI Gateway offers a ZDR (Zero Data Retention) feature, whereby request data is not retained by the vendor for training or other purposes.
Zero Data Retention means that after completing a user's request, the service provider does not retain any input or output data on its servers.
Timeline (last 90 days)
Inference Router offers zero data retention, meaning request and response data are not retained.
Zero data retention usually needs to be enabled separately or meet specific conditions, and is not equivalent to being in effect by default for all requests.
Zero Data Retention (ZDR) is a data governance commitment: after processing a request, the service provider does not persistently store the prompt and model output, nor use them for model training or human review; ZDR generally covers content retention in the inference stage, but metadata such as billing and rate limits is usually still recorded.
Zero data retention means the vendor does not retain customer documents or extraction results in any storage after processing is complete.
Zero-data-retention commitments are usually presented in the form of contracts or terms of service, and the actual implementation varies by vendor: some only refrain from logging at the API level, while others anonymize across the entire infrastructure.
Vercel emphasized MAI's zero-data-retention posture, as well as transparency in training data usage.
When teams choose an AI vendor, they should judge whether a zero-data-retention commitment is sufficient in light of their own compliance requirements (such as GDPR, HIPAA, or internal industry standards), and they remain responsible for what data is sent locally and for their logging strategy.
Zero retention does not equal zero learning—the fact that data is not stored does not mean that patterns and information extracted from it will not influence the model in some form
AI Gateway offers a ZDR (Zero Data Retention) feature, whereby request data is not retained by the vendor for training or other purposes.
Zero data retention is a data processing policy in which the service provider does not retain any input or output data on its servers after completing a user's request.
2 more timeline events
All Facts (12)
Zero retention does not equal zero learning—the fact that data is not stored does not mean that patterns and information extracted from it will not influence the model in some form
75%VerifiedAI Gateway offers a ZDR (Zero Data Retention) feature, whereby request data is not retained by the vendor for training or other purposes.
75%VerifiedZero Data Retention means that after completing a user's request, the service provider does not retain any input or output data on its servers.
75%UnverifiedInference Router offers zero data retention, meaning request and response data are not retained.
50%UnverifiedZero data retention usually needs to be enabled separately or meet specific conditions, and is not equivalent to being in effect by default for all requests.
50%UnverifiedZero Data Retention (ZDR) is a data governance commitment: after processing a request, the service provider does not persistently store the prompt and model output, nor use them for model training or human review; ZDR generally covers content retention in the inference stage, but metadata such as billing and rate limits is usually still recorded.
50%UnverifiedZero data retention means the vendor does not retain customer documents or extraction results in any storage after processing is complete.
50%UnverifiedZero-data-retention commitments are usually presented in the form of contracts or terms of service, and the actual implementation varies by vendor: some only refrain from logging at the API level, while others anonymize across the entire infrastructure.
50%UnverifiedVercel emphasized MAI's zero-data-retention posture, as well as transparency in training data usage.
50%UnverifiedWhen teams choose an AI vendor, they should judge whether a zero-data-retention commitment is sufficient in light of their own compliance requirements (such as GDPR, HIPAA, or internal industry standards), and they remain responsible for what data is sent locally and for their logging strategy.
50%UnverifiedZero data retention is a data processing policy in which the service provider does not retain any input or output data on its servers after completing a user's request.
50%UnverifiedFable因Anthropic要审计每次请求而无法满足零数据保留（ZDR）要求，Opus 5则支持ZDR
50%