Where LLM API zero data retention promises stop
What happened
On October 17, Computing Life published an article drawing on OpenAI and Anthropic policy statements to map where zero data retention (ZDR) promises for LLM APIs apply. It separates "not used for training" from "zero data retention," noting that application state, anti-abuse monitoring, caching and legal preservation can all retain data, and that some endpoints and advanced features may fall outside ZDR. The article recommends auditing the call chain and third-party processors, redacting data before it leaves the network, and keeping a path to switch vendors or run local models.
Written by AI from the coverage · updated 3 hours ago
Coverage
Follow the reports to see the story from different sides.
- Computing Life · Share · Yage大模型 API 的 ZDR 承诺解读：提示词可能在哪些环节被保留
大模型 API 的 ZDR 承诺有适用范围，文章梳理了应用状态、反滥用监控、缓存及法律保全可能保留数据的路径。文章结合 OpenAI、Anthropic 的政策说明，区分不用于训练与零数据保留，并强调部分端点和高级功能可能不受 ZDR 覆盖。工程应对包括检查调用链与第三方处理者、在数据出网前脱敏，以及为供应商切换或本地模型部署保留退路。
Heat over time
Heat now 8·Comparable peak 9（Oct 18 12:00）·Comparable change over 24 hours –
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.