Clarification on data usage for model training — Go and Zen #35102

Description

Context: Go and Zen are cloud-hosted AI gateway services provided by OpenCode (Anomaly Co.). When users send prompts through Go/Zen, the requests pass through OpenCode's proxy servers before reaching the downstream model providers.
What's clear: Both Go and Zen docs explicitly state that downstream providers follow a zero-retention policy and do not use data for model training (with listed exceptions for free models).
What's unclear: There is no statement about what OpenCode itself does with user prompts that pass through its servers. The main privacy policy (https://opencode.ai/legal/privacy-policy) lists:
"Other Identifying Information that You Voluntarily Choose to Provide — such as information included in conversations or prompts that you submit to AI"
as a collected category, with the purpose including:
"Providing, Customizing and Improving the Services"
and under "Improving the Services":
"Improving the Services, including testing, research, internal analytics and product development."
This language is broad enough to potentially cover training internal models, but there's no explicit commitment one way or the other.
The question: Does OpenCode (Anomaly Co.) itself use prompts and conversations routed through Go/Zen for any kind of model training or internal product improvement beyond what's strictly necessary to operate the service? If not, would you consider adding an explicit statement to the Go and Zen docs similar to what's already written about downstream providers?
I think a lot of privacy-conscious users would appreciate clarity here. Thanks!

Plugins

No response

OpenCode version

No response

Steps to reproduce

No response

Screenshot and/or share link

No response

Operating System

No response

Terminal

No response