Trust center
Privacy at Ganivra
Ganivra is a self-serve platform for AI and MCP unit economics. This page explains the data involved in accounts, analytics, and product telemetry.
Accounts and authentication
When you create an account, Supabase processes your email address and authentication data so we can secure your workspace. Ganivra stores workspace membership and telemetry metadata needed to provide the product.
Website analytics
We use Google Analytics to understand aggregate website usage, such as visits and interactions. Google may use cookies or similar technologies according to its own privacy terms.
Product telemetry
Ganivra is designed around execution metadata. Model events may include provider, model, operation, tokens, latency, status, error category, and customer-supplied business identifiers. The telemetry model does not store prompt or response bodies; prompt-level economics use customer-supplied prompt IDs and version labels.
MCP telemetry
MCP events may include server name, tool name, transport, authentication category, latency, outcome, direct cost, bounded counts or flags, and links to other steps in the execution. Customers should not send tool arguments, tool results, OAuth scopes, credentials, workspace IDs, prompts, or responses.
Your choices
You can stop sending telemetry at any time by removing the SDK initialization. Account and workspace deletion requests can be sent through the contact details provided in Ganivra communications.
Last updated: August 15, 2026