Tenant team data is used only to provide the service. Team, run, task, memory, and evidence content may include personal data that a user chooses to provide. OAuth credentials stay in transport headers and are not stored in tool inputs or outputs. Tenant records, memories, and vectors are isolated by verified OAuth identity.
When semantic search is enabled, the text of a memory you store, and each search query, is sent to Cognitum's embeddings service to compute a numeric vector. That service runs on Cognitum infrastructure with a self-hosted model; this service stores only the resulting vector alongside your memory and does not share it with other tenants.
Service data may be retained for longer than 30 days so teams can resume work and review evidence. To request access, correction, or deletion, contact ruv@ruv.net. Data is not sold or used to train general-purpose models.