Skip to content
Ontheia Ontheia Docs

AI Tools (Standard Providers)

Here the administrator defines which AI models should be used for internal system tools.

1. Prompt Optimizer

Section titled “1. Prompt Optimizer”

The Prompt Optimizer automatically improves user requests before they are sent to the actual agent.

  • Configuration: Selection of provider (e.g., OpenAI) and model (e.g., GPT-5).
  • Preset: The installer sets provider/model to the install default (same provider as the example agents); the underlying chain binds the optimize step to the Personal Assistant.
  • Requirement: Since this step serves quality assurance, a powerful model should be chosen here.

2. Context Compression (Rolling Summary)

Section titled “2. Context Compression (Rolling Summary)”

The Summarizer automatically compresses older chat history once the context exceeds a configurable token threshold. This keeps conversations running smoothly even over very long sessions.

  • Configuration: Selection of provider, model, token threshold and minimum plaintext window (default: 20).
  • Preset: The installer sets provider/model to the install default and the token threshold to 8,000 — compression is active out of the box.
  • Recommendation: A capable model improves summary quality; a fast and cost-efficient model reduces latency during compression.

Full description: Context Compression

Note on Selection

Section titled “Note on Selection”

Changes to these providers take effect immediately. Ensure that the selected providers are correctly configured in the “AI Provider” tab and the associated API keys are stored.