Documentation

Meibel developer documentation

Deploy a policy plane in front of your LLM calls in under 15 minutes. Start with the quickstart or browse by topic.

Quickstart guide API reference
Quickstart

Get a policy plane running against a live LLM endpoint in 15 minutes. Covers proxy setup, your first policy config, and verifying redaction output.

Estimated time: 15 min
API Reference

Complete REST API reference. Endpoint listing, authentication, request/response schemas, error codes, and rate limit headers.

REST · JSON · Bearer token auth
Concepts

Architecture overview: tenant isolation model, policy DSL, redaction modes, audit log schema, and the request lifecycle through the policy plane.

Architecture · Glossary
Frequently referenced
Setup
API
Concepts
Five-minute integration

Point any OpenAI-compatible client at your Meibel proxy endpoint. One environment variable change.

Before (direct API call) 
import openai

client = openai.OpenAI(
  api_key="sk-...",
  base_url="https://api.openai.com/v1"
)
After (through Meibel policy plane) 
import openai

client = openai.OpenAI(
  api_key="sk-...",
  # Route through Meibel — all policy controls fire automatically
  base_url="https://proxy.meibelai.org/v1/openai",
  default_headers={"X-Meibel-Tenant": "finance-wealth"}
)