Securing data routing paths, payload volatile limits, and KMS decoupling enclaves.
We route your API requests on the fly. We do not store, look at, or sell the content of your prompts or completions. Everything is processed in-memory and streamed directly to the model provider.
Routor measures performance (tokens, latency, cost) to route your prompts efficiently. We collect metadata like execution speed and token count to generate your cost analytics dashboard.
Your third-party API keys are stored in secure encryption enclaves. We decrypt them only in-memory when sending requests to OpenAI, Anthropic, or other providers.
We route prompts to providers you choose. We contractually require our providers not to use your queries to train their AI models.
Routor is committed to structural open source transparency and developer security. If your enterprise requires customized data containment policies or on-premise orchestration layouts, you can initiate a secure handshake with our core team.