ramen ai
terminal

Developer Hub & Quickstarts

Everything you need to integrate ramen ai into your stack.

1

Installation

Install the SDK via npm to get started.

npm install @ramen-ai/sdk

Then import and initialise the client:

import { RamenClient } from '@ramen-ai/sdk';

const client = new RamenClient({
  apiKey: process.env.RAMEN_API_KEY,
});
2

The Shield — Agent Firewall

Wrap any tool call with withShield to enforce guardrails before execution reaches your downstream services. 

import { withShield } from '@ramen-ai/sdk';

// Define your tool — e.g. a database query
const queryDatabase = async (sql) => {
  return db.execute(sql);
};

// Wrap it with the Shield
const safeDatabaseQuery = withShield(queryDatabase, {
  guardrailId: 'grd_sql_injection_check',
  onBlock:     'throw',   // or 'log' | 'fallback'
  metadata:    { source: 'agent-tool-call' },
});

// Use it like normal — the Shield intercepts automatically
const result = await safeDatabaseQuery(
  'SELECT * FROM users WHERE id = ?'
);

security Why use the Shield?

Every tool call is validated against your configured guardrails before execution. Blocked calls never reach your infrastructure.

3

Standard Integration Guides

Dive deeper into the platform's core features.

shield

System Constraints

Configure master prompts and system-level instructions for your agents.

 gpp_maybe

Guardrails & Guidelines

Set up safety policies, content filters, and compliance rules.

 dns

Deployed Endpoints

Manage your expert agents and their API endpoints.

 verified_user

Compliance Ledger

Review audit trails and compliance reports for all agent activity.
ramen ai