Does Olakai on-prem send our prompts to a cloud AI model?
It is your call: you choose where analysis models run for each task. Olakai on-prem can route analysis to an OpenAI-compatible inference server inside your own network (such as vLLM or Ollama) so nothing leaves your perimeter, or you can route specific analyses to cloud providers under your own API key where Olakai is not in the path.
Self-Hosted Local Inference and Boundaries
You can point analysis at an OpenAI-compatible inference server inside your own network, such as vLLM or Ollama, so nothing leaves your network. If an internal model becomes unavailable, Olakai waits instead of failing over to an external cloud provider.
Source: Your AI Governance Data Is Sensitive Data. Here’s How Olakai Runs Inside Your Walls.
Isolation from Product Infrastructure
Aside from the inference endpoints you designate, the on-prem deployment does not talk to Olakai's product infrastructure, and there is no Stripe, Google Analytics or PostHog traffic. Fully air-gapped setups are supported through internal image mirroring and manual bundle transfer.
Source: Your AI Governance Data Is Sensitive Data. Here’s How Olakai Runs Inside Your Walls.
Also asked as
- Where does prompt analysis run in an Olakai on-prem deployment?
- Can Olakai on-prem process prompts without sending data to cloud AI providers?
- Does Olakai send prompt text to OpenAI or Anthropic when deployed on-premises?