Legal
We built PRISM to simulate AI infrastructure, not to collect your proprietary data. Here is exactly what we collect, what we don't, and how we protect it.
Last updated: March 29, 2026
Unlike production proxies or tracing tools, PRISM is a pre-production simulator. We do not process, store, or transmit your live end-user prompts or AI model completions. Simulations are run using synthetic Monte Carlo traffic generation and empirical behavioral models. Your pipeline configurations never make live calls to third-party model providers (e.g., OpenAI, Anthropic) during a standard simulation.
To provide the PRISM service, we collect the following:
We use your pipeline configurations strictly to:
We will never use your pipeline topologies to train our own models or share your proprietary architectures with competitors.
We use a limited number of trusted third-party infrastructure providers to run PRISM. All subprocessors are bound by strict data processing agreements:
You have the right to access, modify, or permanently delete your pipeline data and account at any time. If you wish to execute a full data deletion request, you can do so directly from your Account Settings or by contacting us at privacy@getprism.dev.