- Is Skyflow an LLM gateway?
- No. Skyflow is a data privacy vault. It stores structured sensitive data and returns tokens to applications. It is not positioned as a gateway that sits in front of LLM providers on the prompt path.
- Does Skyflow mask prompts before model calls?
- Not as a primary capability. Skyflow's model is to tokenise data at storage time so applications hold tokens instead of raw values. Whether a prompt sent to an LLM contains sensitive content depends on what the application assembles before the call.
- Can Skyflow and Privian be used together?
- Yes. A common shape is application → Skyflow (vault for structured sensitive data) → Privian (privacy-first LLM gateway on the prompt path) → LLM provider. The vault protects persistent storage; the gateway protects what reaches the model.
- When do I need a privacy vault?
- When your primary risk is structured sensitive data living in your own systems — customer records, payment data, health information — and you want to centralise tokenisation, access policy and residency controls for that storage.
- When do I need a privacy-first LLM gateway?
- When your primary risk is what reaches an LLM provider in prompts — support tickets, internal documents, user messages — and you want supported entities masked before the provider call and rehydrated in the response.
- What is the difference between tokenisation and prompt masking?
- Tokenisation replaces a sensitive value with an opaque token at storage time; the mapping lives in the vault and the application stores the token. Prompt masking replaces supported entities with deterministic placeholders at request time, calls the model with the masked prompt, and rehydrates the response in-process — nothing about the prompt is persisted.
- Does Privian block prompt injection?
- No. Privian focuses on prompt-level data protection — masking supported personal and sensitive values before they reach the model. It does not claim to detect or block prompt injection or jailbreaks. If injection defence is your primary requirement, a dedicated LLM firewall is a better fit.
- Does Privian support native streaming?
- Not in the current beta. The gateway accepts stream: true and returns artificially chunked text, but it does not pass through native provider token streams yet.