Skip to main content
This page provides an overview of the prompt tuning options available for the GraphRAG indexing engine.

Default prompts

The default prompts are the simplest way to get started with the GraphRAG system. They are designed to work out-of-the-box with minimal configuration. More details about each of the default prompts for indexing and query can be found on the manual tuning page.

Quick start

Use default prompts to get started immediately with minimal configuration

Auto tuning

Auto tuning leverages your input data and LLM interactions to create domain-adapted prompts for the generation of the knowledge graph. It is highly encouraged to run it as it will yield better results when executing an index run.

Auto tuning

Generate domain-adapted prompts automatically from your input data
Auto tuning is highly recommended for better indexing results tailored to your specific domain.

Manual tuning

Manual tuning is an advanced use-case. Most users will want to use the auto tuning feature instead. Details about how to use manual configuration are available in the manual tuning page.

Manual tuning

Customize prompts manually for advanced use cases
Manual tuning is an advanced feature. Most users should use auto tuning instead.

Tuning workflow

1

Initialize your workspace

Run graphrag init to create the necessary configuration files and default prompts.
2

Choose your tuning method

Decide between auto tuning (recommended) or manual tuning based on your needs.
3

Run prompt tuning

Execute the prompt tuning process using the CLI or by manually editing prompt files.
4

Update configuration

Modify your settings.yaml to reference the new prompts.
5

Run indexing

Execute graphrag index with your tuned prompts for better results.

Next steps

Auto tuning

Learn how to automatically generate domain-adapted prompts

Manual tuning

Explore advanced manual prompt customization

Configuration

Configure GraphRAG settings for optimal performance

Indexing

Run the indexing pipeline with your tuned prompts

Build docs developers (and LLMs) love