Skip to content

Obstruo

Obstruo is a lightweight AI Governance Control Plane. It enforces policy on the LLMs and agents your teams already run, whoever built them. Sensitive data is masked before a prompt leaves your infrastructure, unsafe requests are blocked, and every call is written to an audit trail.

Obstruo exposes an OpenAI-compatible API. If your code already talks to OpenAI, switching to Obstruo is a two-line change:

import openai

client = openai.OpenAI(
    api_key="your-obstruo-key",
    base_url="https://api.obstruo.ai/your-org/v1",
)

response = client.chat.completions.create(
    model="gpt-4o-prod",   # a Logical Model you configured in the panel
    messages=[{"role": "user", "content": "Hello!"}],
)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "your-obstruo-key",
  baseURL: "https://api.obstruo.ai/your-org/v1",
});

const response = await client.chat.completions.create({
  model: "gpt-4o-prod",
  messages: [{ role: "user", content: "Hello!" }],
});

Obstruo is not tied to any vendor or cloud ecosystem. The same rules apply whether a request goes to OpenAI, Anthropic, Azure, or a model you host yourself.

Every call that passes through the gateway gets the full pipeline. Personal and sensitive data is detected and masked before the prompt reaches an LLM. Prompt injection, jailbreak attempts, and unsafe content are screened in both directions, with rules configurable per model, tool, client, or risk level. Latency, errors, and cost are tracked along the way, and a tamper-evident audit record is written for every request and response.

What's in these docs