Setup

Guardrails on your agent in three steps

No rewrite. Install the SDK, wrap your agent or point your client at the gateway, then set your limits.

Steps

  1. 1

    Install

    pip install steerplane or npm install steerplane.

  2. 2

    Wrap or proxy

    Add @guard to your agent, or change one base_url.

  3. 3

    Set your limits

    Cost, steps and denied actions. In code or a YAML file.

agent.py
from steerplane import guard

@guard(
    agent_name="support_bot",
    max_cost_usd=5.00,
    max_steps=50,
    denied_actions=["delete_*", "drop_*"],
)
def run_support_agent():
    agent.run()  # your code, unchanged
Two ways in

Pick the path that fits your stack

Decorator

Wrap the function that runs your agent. Best when you own the agent code and want the deepest control.

Gateway

Change your OpenAI client's base_url. No code changes inside the agent. Your provider key stays on the server.

Integrations

Works with what you already use

LangChainCrewAIAutoGenOpenAI Agents SDKPlain PythonTypeScript / Node.js
Self-host

Run the full stack on your own servers

The API, gateway, dashboard and database come up with one command.

git clone https://github.com/vijaym2k6/SteerPlane.git
cd SteerPlane
cp .env.example .env
docker compose up -d   # API, gateway, dashboard and Postgres
Read the docsView on GitHub
Design partners

Shipping agents? Let's talk.

Our first 10 design partners get every paid feature free, plus direct help setting up.

Book a callGet access