How Trovella's AI Works

Overview

Trovella uses large language models (LLMs) to plan, execute, and synthesize research. This page explains what happens behind the scenes when you create a research plan.

Plan Creation

When you submit a research question, Trovella's AI analyzes it and creates a structured plan. The plan breaks your question into discrete steps, each targeting a specific aspect of your query.

Step Execution

Each step in the plan is executed independently. The AI searches for relevant sources, evaluates their credibility, and extracts key findings. Sources are tracked so you can always verify claims.

Synthesis

After all steps complete, the AI synthesizes findings into a coherent summary. It identifies patterns, contradictions, and gaps in the research.

AI Models

Trovella uses Anthropic's Claude models for reasoning and Google's models for specialized tasks. All AI usage is logged and visible in the admin dashboard.