Radzor Component Spec v1.0.0

Components
for the AI era

A universal registry of standardized building blocks with machine-readable manifests that LLMs can discover, understand, and compose together.

radzor.manifest.json
{
  "radzor": "1.0.0",
  "name": "@radzor/semantic-search",
  "inputs": [{ "name": "query", "type": "string" }],
  "actions": [{ "name": "search" }]
}

Why Manifest?

A standardized way to describe AI components so any LLM can understand and use them.

Universal Standard

One manifest format that works across all LLM providers, frameworks, and languages.

Instant Integration

Import a manifest and your agent immediately knows how to interact with the component's API and logic.

Type-Safe AI

Strict schemas for inputs and outputs prevent hallucinations and ensure data integrity in your AI workflows.

Available Components

Standardized building blocks ready for LLM-driven development.

Browse Registry
Audio

Audio Capture

Browser-based audio recording with configurable sample rate, format, and real-time streaming.

Auth

OAuth Authentication

Multi-provider OAuth 2.0 authentication with session management and token refresh.

Payment

Stripe Checkout

Complete Stripe payment integration with subscriptions, webhooks, and refund handling.

Chat

Realtime Chat

WebSocket-based real-time messaging with rooms, typing indicators, and message history.

The Manifest Standard

We're proposing a universal format for AI components. A Manifest file describes not just the API, but the "soul" of the component—how it should be prompted, its constraints, and its intended behavior.

  • Self-documenting capabilities for LLMs
  • Standardized tool-calling definitions
  • LLM-facing constraints and usage guidelines
  • Versioned and immutable definitions
Read the Specification

Integration Example

manifest-sdk-v1.0

EXAMPLE
// 1. Fetch the manifest from the registry
const manifest = await fetch("radzor.io/api/components/audio-capture/manifest");
// 2. LLM reads inputs, outputs, actions
const { inputs, actions } = manifest;
// 3. LLM generates code using the spec
await llm.generate("Add audio recording to my app", { manifest });

Ready to build the
future of AI?

Start building AI-native applications with standardized components.