AI Scraping terms — is per-user summarization of a user's own channel feed allowed for a TDLib client?

Setup: third-party TDLib client, own api_id, following the client requirements in the API ToS.

Feature: each user summarizes their own subscribed channels on demand — per-user auth, own feed only, run on Cocoon (never leaves the ecosystem), output shown only to that user, nothing stored, pooled, or republished. Effectively your first-party AI summaries in a compliant client.

Question: is this legitimate client use under the API ToS, or does it need the consent/exception under the Content Licensing & AI Scraping terms (§1.5)?

Thanks!