Data governanceunder review
What happens to your data: training-on-customer-data defaults, retention windows, residency options, and the enterprise-versus-consumer terms gap. A legal property, not a capability — it survives every model generation.
Stays under-review, on firmer grounds than before. Tencent Cloud does publish a general privacy policy, and it was read in full: across roughly 92,000 characters it does not mention Hunyuan, model training, generative AI, or large language models even once. So there is no first-party statement — in either direction — on whether prompts submitted to the TokenHub API are used for training, and no Hunyuan-scoped retention or residency terms exist to cite. This is an absence of published terms, not an unread page. The Apache 2.0 weights make self-hosting fully deployer-controlled.
Receipts (2)
- Tencent Cloud Privacy Policy (general; silent on model training and Hunyuan)
“We will retain your personal information as described below (except if otherwise required to be retained by applicable law).”
Quoted to establish what this policy does cover. The finding is what it omits — an absence cannot itself be quoted, which is why this vector stays under-review rather than being graded from silence. - Tencent Cloud TokenHub product page (pricing and plans; no data-handling terms)Retained to record that the URL this registry previously cited as data-handling evidence is a marketing and pricing page containing none.