Proxide Documentation

Proxide is a drop-in AI gateway proxy. Change your baseURL and get automatic failover, per-agent budget limits, PII redaction, semantic caching, and loop detection — without changing anything else.

One-line setup

TypeScript / JavaScript
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "prox-your-key-here",
  baseURL: "https://gateway.proxide.ai/openai/v1",
});

// Everything else stays the same

Works with any OpenAI-compatible client: Python SDK, JS SDK, curl, LangChain, LlamaIndex, and more.

Documentation

Quick start

Sign up, get your API key, and change your baseURL. Working in under 2 minutes.

Budget limits

Set per-agent daily and monthly spend limits. Returns 402 on exceed. Handle the error gracefully.

PII redaction

Automatic detection and redaction of emails, card numbers, SSNs, NI numbers, and more.

Semantic cache

Vector-similarity caching for LLM responses. Cache hits served in <5ms. Saves 20–40% on API costs.

Failover & routing

Configure provider fallback chains. Automatic retry on 429s, 5xx errors, and timeouts.

Gateway endpoints

OpenAI-compatiblehttps://gateway.proxide.ai/openai/v1
Anthropic-compatiblehttps://gateway.proxide.ai/anthropic/v1