Skip to content

Prompts Overview

Introduction

EcoSemantic integrates with AI assistants like Claude through the Model Context Protocol (MCP), allowing you to perform Life Cycle Assessment (LCA) calculations using natural language. This section will teach you how to effectively communicate with your AI assistant to leverage EcoSemantic's powerful tools.

What Are Prompts?

Prompts are the natural language instructions you give to your AI assistant. With EcoSemantic connected, your AI assistant can:

  • Search the Ecoinvent database for activities
  • Calculate environmental impacts
  • Explore impact methods and data structures
  • Create custom LCA projects
  • Perform advanced analyses

Core Capabilities

🔍 Searching Activities

Find processes in the Ecoinvent database using descriptive keywords.

Example: 

"Search for electricity production activities in Germany"

📊 Calculating Impacts

Calculate environmental impacts for specific activities and quantities.

Example: 

"Calculate the carbon footprint of 100 kWh of solar electricity from the US"

🌍 Exploring Data

Discover available impact methods, units, and biosphere compartments.

Example: 

"What impact methods are available for water scarcity?"

🏗️ Custom Projects

Build your own LCA models for technologies not in Ecoinvent.

Example: 

"Create a project for hydrogen fuel cell vehicle LCA"

How to Use This Section

This prompts guide is organized by complexity:

Basic Prompts

Start here if you're new to EcoSemantic:

Advanced Prompts

More advanced capabilities will be documented soon. For now, explore the Examples page for complex workflows.

Writing Effective Prompts

Be Specific

Instead of: "Find electricity"
Try: "Search for wind electricity production in Denmark"

Provide Context

Instead of: "Calculate carbon footprint"
Try: "Calculate the carbon footprint of 1000 kg of steel production using IPCC 2013"

Ask for Clarification

If results aren't what you expected, ask follow-up questions: 

"Show me only activities from the US"
"Can you filter for processes with 'natural gas' in the name?"
"What's the difference between these two activities?"

Chain Requests

Break complex analyses into steps: 

1. "Search for cement production activities"
2. "Calculate carbon footprint for the top 3 results"
3. "Compare the results in a table"

Understanding Responses

When you make a request, your AI assistant will:

  1. Use EcoSemantic tools - Call the appropriate MCP tools
  2. Process results - Interpret the data returned
  3. Present findings - Format the information clearly
  4. Suggest next steps - Recommend follow-up actions

Pro Tip

The AI assistant can see which tools it has available. You can ask: "What EcoSemantic tools do you have access to?"

Common Patterns

Pattern 1: Search → Calculate → Analyze

"Search for aluminum production activities, 
calculate their carbon footprints, 
and tell me which has the lowest impact"

Pattern 2: Explore → Select → Deep Dive

"What impact methods are available for climate change?
Use IPCC 2013 to calculate impacts for natural gas electricity"

Pattern 3: Create → Build → Validate

"Create a project for electric vehicle battery manufacturing,
add activities for cathode and anode production,
validate the exchanges are complete"

Available Databases

EcoSemantic supports multiple Ecoinvent versions:

Database Description Best For
ecoinvent-3.11-cutoff Latest version (2024) Most current data
ecoinvent-3.10.1-cutoff Previous stable Compatibility
ecoinvent-3.9.1-cutoff Widely used Research reproducibility
ecoinvent-3.8-cutoff Stable version Legacy projects
ecoinvent-3.7.1-cutoff Older stable Historical analysis
ecoinvent-3.6-cutoff Legacy Specific requirements

Database Selection

Unless you have specific requirements, use ecoinvent-3.11-cutoff for the most recent data.

Next Steps

Ready to get started? Begin with:

  1. Searching Activities - Learn to find processes
  2. Calculating Impacts - Run your first LCA calculation
  3. Examples - See real-world use cases

Need Help?

Let's begin! Head to Searching Activities to learn your first prompts.