Getting Started

The front door is the Pulse Gateway: an OpenAI- and Anthropic-compatible endpoint. Change the base URL of the SDK you already use and every call is PII-masked before egress, routed to the right model, and attested by a signed certificate a third party can verify offline. Start with the five-minute quick start.

Behind the gateway sit two engines that run on top of whatever AI you use: Synapse decides how to call a model — routing, PII redaction, cost tracking across any provider — and Cortex decides what the model knows, compiling your documents into a cited, verified knowledge layer. The platform APIs below ride on both.

Pulse Gateway

The attested endpoint →

Synapse Engine

How models are called →

Pulse Cortex

What models know →

Base URL

production

https://api.pulse-labs.fr

local dev

http://localhost:3051

Quick start

bash

# The fastest path: point the SDK you already use at the gateway.
# One changed line — every call PII-masked, routed, and attested.

curl -s https://api.pulse-labs.fr/v1/chat/completions \
  -H "Authorization: Bearer sk_live_your_key_here" \
  -H "Content-Type: application/json" \
  -D - \
  -d '{"model": "pulse-auto", "messages": [{"role": "user", "content": "Hello"}]}'

# ← the x-pulse-proof response header is the signed certificate of this call.
# Full walkthrough: /docs/quickstart