> For the complete documentation index, see [llms.txt](https://aro-1.gitbook.io/aro/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://aro-1.gitbook.io/aro/aros-tech/data-flow-and-multimodel-reasoning.md).

# Data Flow & Multimodel Reasoning

<figure><img src="/files/qLiqgPCflacQjj3vrJIh" alt=""><figcaption></figcaption></figure>

**User Request**

* Example: “What is the long-term analysis for Bitcoin $BTC”
* Sent via  **chat** , **UI dashboard, API call.**

**AI Research Orchestrator Receives the Task**

* Classifies the query (market sentiment, price prediction, or code interpretation).

**Multimodel Reasoning Layer**

* **Model Selection**: Chooses o1/deepseek r1 for textual reasoning, or Grok for advanced sentiment analysis, etc.
* **Collaboration**: Multiple LLMs can be used concurrently for complex, multi-part questions.

**Tool Interactions**

* The orchestrator invokes relevant tools in the **Orchestration Intelligence** (e.g., on-chain, sentiment, or financial analytics) for data gathering or specialized computations.

**Aggregation & Finalization**

* Partial outputs (from LLMs + Orchestration Intelligence) are **merged** into a cohesive final answer.
* Logic checks by AI Research Orchestrator ensure data consistency and integrity.

**Dataset Logging**

* Key outputs are **archived** for historical reference and to refine future model performance.

**Output Delivery**

* Final insights appear in web app chat reply, an interactive dashboard, PDF/HTML report, or as a structured **API** response.