Custom Reflections

Overview

Reflections are AI-generated interpretive texts that add depth to your signals, clusters, and shapes. Autonomy Realms includes four built-in reflection types, but you can define your own custom reflection types tailored to your realm's needs.

Entity Levels

Custom reflections can be defined at three entity levels:

• Signal — Reflections generated for individual signals
• Cluster — Reflections generated for groups of related signals
• Shape — Reflections generated for geographic shapes (areas of interest)

Each entity level has its own set of custom reflection definitions. A reflection type defined at the signal level will only appear when generating reflections for individual signals.

Defining a Custom Reflection

Each custom reflection type requires:

• Name — A label for the reflection type (e.g., "Ecological Note" or "Historical Context")
• System Prompt — Instructions that tell the AI how to behave and what role to adopt
• User Prompt — The specific request sent to the AI, including placeholders for signal data

Default Prompts

When you create a new custom reflection, the system pre-fills the System Prompt and User Prompt fields with sensible defaults. These defaults serve as starting points — you can use them as-is or modify them to suit your specific needs.

The default prompts demonstrate the expected format and show which data placeholders are available for inclusion in your prompts.

Use Cases

Custom reflections let you extend the AI's interpretive capabilities in directions specific to your archive. Examples include:

• A botanical realm adding a "Species Identification" reflection
• A historical archive adding a "Period Context" reflection
• A sound art realm adding a "Acoustic Analysis" reflection

Each custom type generates its own distinct output, appearing alongside the built-in reflections in the signal detail view.