> For the complete documentation index, see [llms.txt](https://asknoel.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://asknoel.gitbook.io/docs/asknoels-tech/processing-pipeline.md).

# Processing Pipeline

#### <mark style="color:purple;">**User Request Handling**</mark>

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

#### <mark style="color:purple;">**ASKNOEL AI Research System**</mark>

* Identifies query type: **market sentiment, price prediction, or code interpretation**.

<mark style="color:purple;">**Multimodel Reasoning Layer**</mark>

* **Model Selection:** Picks the best model for the task (e.g., DeepSeek for textual reasoning, Grok for sentiment analysis).
* **Collaboration:** Multiple LLMs collaborate for complex, multi-part queries.

<mark style="color:purple;">**Tool Interactions**</mark>

* **ASKNOEL invokes specialized tools** (on-chain analytics, sentiment analysis, financial modeling).

**Aggregation & Finalization**

* Combines results from LLMs and ASKNOEL intelligence tools into a **cohesive final answer**.
* Runs **logic and consistency checks**.

<mark style="color:purple;">**Dataset Logging**</mark>

* Archives key outputs for **historical reference and model refinement**.

<mark style="color:purple;">**Output Delivery**</mark>

* Final insights are delivered via:
  * **Chat replies** in web app
  * **Interactive dashboard**
  * **PDF/HTML reports**
  * **Structured API responses**