Engines
An engine is the core system that checks decisions in Averment.
If you think of it like LLMs: Engine = Model
Unlike LLMs, Averment engines behave consistently and evaluate actions using multiple layers of reasoning. You choose the engine that decides how actions are evaluated.
Available Engines
Averment is currently in closed beta. The supported engine is:
dde-testHow Averment compares
| Dimension | Comparison |
|---|---|
| Determinism | Deterministic System > Averment > Pure LLM |
| Semantic Intelligence | Pure LLM > Averment > Deterministic System |
| Adaptability | Pure LLM > Averment > Deterministic System |
| Execution Guarantees | Deterministic System > Averment > Pure LLM |
| Operational Safety | Averment > Deterministic System > Pure LLM |
| Contextual Flexibility | Pure LLM > Averment > Deterministic System |
| Bounded Execution Behavior | Averment > Deterministic System > Pure LLM |
| Handling Complex Semantics | Pure LLM > Averment > Deterministic System |
| Trustworthy Autonomous Execution | Averment > Deterministic System > Pure LLM |
Averment is a control layer, not a replacement for fraud or ML systems. It sits above them:
Using an Engine
At initialization
js
Per request
js
If both are provided, the request-level engine takes priority.
Next steps
Now that you understand engines, learn how to customize risk evaluation for your specific domain.
Behavior Intelligence
Configure risk tolerance, semantic overrides, and enforcement rules tailored to your domain without writing code.