What Vellum does with your messages
Read from its published policy on 2026-09-06, and quoted rather than paraphrased. It says it does not train on your conversations. It does not say which company actually carries your messages, which puts it with the majority: only 9 of 41 do.
What the document says
| Trains on your conversations | States it does not |
|---|---|
| Data of people who never signed up | Not addressed |
| If you cancel | Not addressed |
| Model providers named | Anthropic, OpenAI, Google, ElevenLabs |
| Who carries the messages | Not disclosed |
| Where to report a hole | No route published |
| Jurisdiction | Not stated |
In its own words
“but do not use it to train AI models”
What it does not answer
The policy is silent on data of people who never signed up, if you cancel, who carries the messages, where to report a hole and jurisdiction. Silence is not a denial, and it is not a promise either: a document that does not address the people in your group thread has not agreed to anything about them.
Back to the Vellum listing, or see every policy in this directory read the same way, where 9 of 41 name who holds your messages and the rest do not.