- Home
- MCP servers
- MCP Prompt Engine Server
MCP Prompt Engine Server
- go
16
GitHub Stars
go
Language
6 months ago
First Indexed
2 months ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
Installation
Add the following to your MCP client configuration file.
Configuration
View docs{
"mcpServers": {
"vasayxtx-mcp-prompt-engine": {
"command": "mcp-prompt-engine",
"args": [
"serve"
]
}
}
}You can manage and serve dynamic prompt templates with the MCP Prompt Engine. It lets you create reusable, logic-driven prompts using Go templates, expose them as MCP prompt arguments, and serve them to compatible MCP clients with hot-reload and a rich CLI for development and testing.
How to use
Run the server locally or in a container to serve your prompt templates to MCP clients. Create prompts in a dedicated directory, validate them, and then start the server so clients can fetch and render prompts with dynamic input.
How to install
Prerequisites you need before starting:
-
Go is required to install the MCP Prompt Engine as a binary.
-
Docker is optional if you plan to run the containerized server.
# Install the MCP Prompt Engine binary
go install github.com/vasayxtx/mcp-prompt-engine@latest
# Create a prompts directory and add templates (see Usage for template structure)
mkdir -p prompts
# Add your .tmpl files here
Run the server locally
Start the server so your prompt templates are available to MCP clients. By default, the server looks for templates in the ./prompts directory.
mcp-prompt-engine serve
To customize the prompts directory or enable quiet mode, pass explicit flags as shown in the following variations.
## Run the server with Docker
If you prefer containerized deployment, run the pre-built Docker image. Mount your local prompts and logs into the container for persistence.
Pull and run the pre-built image from GHCR
docker run -i --rm
-v /path/to/your/prompts:/app/prompts:ro
-v /path/to/your/logs:/app/logs
ghcr.io/vasayxtx/mcp-prompt-engine
## Validate and test prompts
Validate prompts to catch syntax errors before starting the server. Use the CLI to list, render, and validate templates during development.
Validate a specific prompt
mcp-prompt-engine validate git_stage_commit
List prompts with details
mcp-prompt-engine list --verbose
Render a prompt with arguments (example)
mcp-prompt-engine render git_stage_commit --arg type=feat
## Connecting to clients
Your MCP client will discover and render prompts exposed by this server. Ensure the client is configured to connect to your MCP server entry, which can be the local binary or the Docker container endpoint, depending on how you deployed.
## Notes and tips
- The server automatically parses JSON arguments when possible, enabling complex data types for template inputs. You can enable or disable JSON parsing with a flag if needed.
- Partial templates prefixed with an underscore can be included in other templates to promote reuse.
- The Go text/template engine powers all templating features, including variables, conditionals, loops, and partials.
- For development, hot-reload detects changes to prompts and reloads without restarting the server.
## Troubleshooting
If prompts are not appearing in clients, ensure the server is running, the correct prompts directory is mounted, and there are no syntax errors in the templates.
## Available tools
### validate
Checks all templates for syntax errors and reports any issues.
### list
Lists available prompts and their descriptions; with --verbose shows detailed info and variables.
### render
Renders a template directly in the terminal using provided arguments and environment fallbacks.
### serve
Starts the MCP server to expose prompts to MCP clients.