Skip to content
FastPII
AI Gateway

Route every LLM request through a security checkpoint

FastPII exposes an OpenAI-compatible proxy at /v1/chat/completions so engineering teams can stream responses, bring their own keys, and enforce central decisions before any model sees raw traffic.

Start FreeView Documentation

Request path

App → Gateway → Policy → Provider → Audit

Streaming

Server-Sent Events keep token streams intact.

Providers

OpenAI, Azure OpenAI, Anthropic, and Ollama.

Gateway capabilities

Built for live AI engineering traffic

The gateway keeps the familiar chat completions contract while adding inspection, routing, and governance controls around every request.

OpenAI-compatible endpoint

Drop into existing chat pipelines with the /v1/chat/completions shape your tooling already expects.

Streaming over SSE

Keep interactive response streaming intact while policies run on the request path.

BYOK routing

Store provider credentials centrally and keep model access tied to your organisation’s own keys.

Multi-provider dispatch

Route the same protected request shape to OpenAI, Azure OpenAI, Anthropic, or Ollama.

Policy enforcement

Seven actions, one central decision layer

Policies can stop, transform, flag, or route traffic based on what the Detection Engine found and which workspace or provider is involved.

ALLOW

WARN

MASK

BLOCK

ESCALATE

LOG

QUARANTINE

Code example

Keep the chat API your team already knows

Teams swap the endpoint, add FastPII credentials, and let the gateway handle secure forwarding and audit logging behind the same request contract.

Provider coverage

  • • OpenAI
  • • Azure OpenAI
  • • Anthropic
  • • Ollama
The gateway forwards streaming responses, preserves BYOK boundaries, and records decisions for later audit export.
/v1/chat/completions
curl https://api.fastpii.com/v1/chat/completions \
  -H "Authorization: Bearer $FASTPII_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4.1-mini",
    "stream": true,
    "messages": [
      {"role": "user", "content": "Review customer note: PESEL 90010112345"}
    ]
  }'
Deploy the checkpoint

Put FastPII in front of every model call

Use the gateway when you need one place to enforce inspection, provider controls, policy outcomes, and audit evidence across AI workloads.

Start FreeView Documentation