PRD Creator MCP Server

How to use

You connect an MCP client to this server to generate PRDs from product descriptions, user stories, and requirements. Choose your AI provider or a template-based path, configure provider options, and request PRD creation. If an AI provider cannot respond in time, the server gracefully falls back to a template-based PRD.

How to install

Prerequisites: Node.js v16 or higher, npm or yarn.

Install from source by following these steps.

git clone https://github.com/Saml1211/prd-mcp-server.git
cd prd-mcp-server
```
```bash
npm install
```
```bash
npm run build
```
```bash
npm start
```
```bash
npm run dev

Configuration, providers & runtime behavior

Configure AI providers via a .env file or at runtime using protocol tools. You can place a .env file in your project or working directory and populate standard provider variables (for example, keys and model preferences). You can also update provider configuration on the fly with the update_provider_config MCP tool. The server merges persistent configuration with environment variables, giving precedence to updates made through the MCP tools.

At startup you can also validate connectivity and current provider availability using the system health checks and logs commands.

Using via help and basic usage

Get help and see usage options by running the server with the help flag or using its CLI entry point. Typical entry points are available via npx, or a globally installed CLI if you choose to install it.

You can initiate PRD generation, validation, and template management using the provided tools.

Integrations & runtime helpers

Integrations describe how to connect with client tools and desktop environments that drive MCP workflows. Examples show how to configure clients to launch the MCP server using npx or other runtime commands.

Notes on templates and templates management

The server provides a library of PRD templates and supports creating, listing, getting, updating, deleting, exporting, and importing templates. You can render a template with concrete placeholders to tailor it to a product context.

Common commands at a glance

Run the MCP server with NPX for a quick start: npx -y prd-creator-mcp. You can also run it via Docker or use the CLI if you install it globally.

generate_prd

Generate a complete PRD using AI-driven or template-based generation, including product name, description, target audience, core features, constraints, and provider options.

validate_prd

Validate a PRD against best practices and customizable rules to ensure completeness and quality.

list_validation_rules

List all available validation rules to understand the criteria used for PRD validation.

list_ai_providers

List available AI providers and their current availability status to help you choose generation options.

create_template

Create a new PRD template to support different product types.

list_templates

List all available PRD templates.

get_template

Retrieve a specific PRD template by name.

update_template

Update an existing PRD template with new content or structure.

delete_template

Delete a PRD template no longer needed.

export_templates

Export all templates to a JSON file for backup or sharing.

import_templates

Import templates from a JSON file into the server.

render_template

Render a template by filling in placeholders with provided context.

get_provider_config

Retrieve the current AI provider configuration.

update_provider_config

Update provider credentials and models at runtime without restarting the server.

health_check

Check system health and provider availability to monitor readiness.

get_logs

Fetch recent system logs for debugging and auditing.

stats

Get usage statistics for monitoring and reporting.