> For the complete documentation index, see [llms.txt](https://gitbook.edgen.tech/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://gitbook.edgen.tech/edgen-litepaper/technical-principles/how-edgm-thinks.md).

# How EDGM Thinks

The same structure that routes crypto and stock queries through agents and tools can also be viewed as a layered thought process. [EDGM](/edgen-litepaper/technical-principles/edgm-the-commander-model.md) assigns tasks, but goes a step further too: it orchestrates them, shaping raw inputs into a single line of reasoning.

[EDGM](/edgen-litepaper/technical-principles/edgm-the-commander-model.md) is designed to think the way an organized research desk would, but at machine scale. Instead of one giant model guessing at answers, it acts as a commander that coordinates layers of analysis.

<figure><img src="/files/dfwzMR7ed8wevS8VaoSP" alt=""><figcaption></figcaption></figure>

It begins with a **personal engine**. This engine interprets the user’s intent, adapts to their portfolio or query, and decides which type of analysis to activate.

It then calls on **specialist agents**. Each agent is narrow by design: fundamentals, technicals, momentum, macro, or sector. Their focus makes outputs precise. Together, they cover the different angles needed to understand a market move.

These agents rely on **market tools:** earnings reports, derivatives data, on-chain activity, macro releases. Outputs are collected, cleaned, and cross-checked against Edgen’s knowledge base.

**Finally, the commander fuses everything into a single answer.** Instead of fragmented views from separate tools, the output is coherent, connected, and decision-ready.

This layered process delivers three things professional traders demand:

* **Coherence:** one answer that combines all relevant perspectives.
* **Adaptability:** when conditions shift in one domain, agents in that domain adapt without breaking the system.
* **Efficiency:** compute is allocated only where needed, making responses fast and cost-effective.

The result is a system that thinks more like an experienced team of analysts than a generic AI, but runs continuously and at global scale.

As we mentioned, [EDGM](/edgen-litepaper/technical-principles/edgm-the-commander-model.md) works like a head chef in a professional kitchen. It doesn’t cook everything itself. It directs specialists (one prepares proteins, another handles sauces, another manages timing) and brings them together into a single dish. The output is coherent and consistent, not a random mix of parts.