GoReal-AI/plp: Open protocol for prompt library interchange. A standardized REST API for managing and versioning AI prompts across platforms.

AI prompts are hardcoded everywhere:

// ❌ Scattered across your codebase
const prompt = "You are a helpful assistant that generates marketing copy...";
await llm.complete(prompt);

This creates:

• Version chaos – Which prompt version is in production?
• No collaboration – Engineers and prompt experts can’t work independently
• Deployment hell – Changing a prompt requires code deployment
• Zero reusability – Every app reinvents prompt management

PLP is a universal protocol that decouples prompts from code, just like APIs decouple frontends from backends.

// ✅ Clean, versioned, traceable
import { PLPClient } from '@goreal-ai/plp-client';

const plp = new PLPClient('https://prompts.goreal.ai');
const prompt = await plp.get('marketing/welcome-email', '1.2.0');
await llm.complete(prompt.content);

1. RESTful & Simple – Three endpoints: GET, PUT, DELETE
2. Language Agnostic – Works with any stack (Python, JS, Go, Rust…)
3. Version Control Built-In – Fetch latest or pin to specific versions
4. Open Standard – MIT licensed, community-driven

PLP is an open specification – like HTTP or OAuth. Any prompt management tool can implement it, and any AI framework can consume it.

For Prompt Library Providers: Implement PLP and your service works with every PLP client.

For AI Framework Developers: Add a PLP client and your framework works with any PLP-compliant prompt library.

Learn more: Protocol Overview | Full Specification

# JavaScript/Node.js
npm install @goreal-ai/plp-client

# Python
pip install plp-client

JavaScript Example:

import { PLPClient } from '@goreal-ai/plp-client';

const client = new PLPClient('https://your-plp-server.com', {
  apiKey: 'optional-key'
});

// Get latest version
const prompt = await client.get('product/feature-announcement');
console.log(prompt.content); // "Hello {{name}}..."

// Save a new prompt
await client.put('support/faq-response', {
  content: 'Answer: {{answer}}',
  meta: { author: 'shimon', version: '1.0.0' }
});

// Delete a prompt
await client.delete('deprecated/old-prompt');

Python Example:

from plp_client import PLPClient

client = PLPClient("https://your-plp-server.com", api_key="optional-key")

# Get latest version
prompt = client.get("product/feature-announcement")
print(prompt.content)  # "Hello {{name}}..."

# Save a new prompt
client.put("support/faq-response", {
    "content": "Answer: {{answer}}",
    "meta": {"author": "shimon", "version": "1.0.0"}
})

Every prompt follows this structure:

{
  "id": "marketing/welcome-email",
  "content": "Hello {{name}}, welcome to {{product}}!",
  "meta": {
    "version": "1.2.0",
    "author": "shimon",
    "description": "Sent on user signup",
    "model_config": { "temperature": 0.7 }
  }
}

See Full Specification for details.

npm install @goreal-ai/plp-express-middleware

import express from 'express';
import { plpMiddleware } from '@goreal-ai/plp-express-middleware';

const app = express();

// Add PLP endpoints with file-based storage
app.use('/v1', plpMiddleware({
  storage: './prompts-db'
}));

app.listen(3000);
// Now serving: GET/PUT/DELETE /v1/prompts/{id}

Use our OpenAPI spec to generate server stubs:

# Generate Go server
openapi-generator generate -i spec/openapi.yaml -g go-server

# Generate Python FastAPI
openapi-generator generate -i spec/openapi.yaml -g python-fastapi

Or manually implement the three endpoints.

Want to build an SDK? Follow the Implementation Guide.

Want to build a PLP-compliant server or client?

1. Read the Full Specification
2. Follow the Compliance Guide
3. Test your implementation
4. Add the compliance badge:

[![PLP Compliant](https://img.shields.io/badge/PLP-v1.0_Compliant-brightgreen)](https://github.com/gorealai/plp)

• GitHub Discussions – Ask questions, share implementations
• Contributing – See CONTRIBUTING.md for guidelines
• Roadmap – Check Issues for planned features

Built with ❤️ by GoReal.AI

MIT © GoReal.AI

See LICENSE for details.