> For the complete documentation index, see [llms.txt](https://docs.aigent.run/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.aigent.run/ai-agent-terminal/agent-frameworks-and-aigentos.md).

# Agent Frameworks & aigentOS

<figure><img src="/files/qJbBPiVpeVubrY34y8J6" alt=""><figcaption><p>Breakdown of aigentOS framework</p></figcaption></figure>

The AI Agent Terminal needs more than just tools, it needs a framework that can understand requests, manage context, and turn them into meaningful actions.

This is the role of **aigentOS**, our proprietary framework built specifically for conversational, terminal-first interaction.

### **What is aigentOS?**

aigentOS is the **core intelligence layer of the terminal**.\
It is lightweight, composable, and optimized for speed, ensuring that every user can interact with powerful agents through simple chat or voice.

aigentOS handles:

* **Understanding your request** → “Swap 589 XRP into RLUSD.”
* **Choosing the right tools** → Liquidity analysis, slippage check, trade execution.
* **Delivering clear results** → “Swap executed at 0.6% slippage, 10,000 RLUSD received.”

### **Context Sources**

Agents powered by aigentOS learn and act based on diverse inputs:

* **Onchain Token Data**: Real-time liquidity, balances, trades, and markets.
* **Wallet History**: Past activity helps personalize decisions.
* **Social Content**: Public feeds (like X/Telegram/Discord) inform analysis.
* **User Documents**: Upload files (PDF, CSV, JSON, and text) for agent reference.
* **Visual Media**: Image/video input support (planned).

This multi-source context ensures that agents are **aware, adaptive, and informed**.

***

### **Vector Storage**

At the heart of aigentOS is a **high-performance vector database**:

* Efficiently embeds and stores context.
* Enables rapid retrieval of relevant data.
* Provides similarity search and clustering for long-term memory.

This allows agents to build continuity across sessions and deliver **smarter responses over time**.

***

### **Active Functions**

Once context is set, aigentOS converts it into targeted actions:

* **Trading Functions**: Analyze markets, calculate slippage, execute swaps.
* **Account Functions**: Summarize balances, track history, bridge assets.
* **Chat Responses**: Generate context-aware, human-like replies.
* **Content Generation**: Create text (with future expansion into visuals).

***

### **Output Composer**

The **Output Composer** is where it all comes together:

* Pulls relevant context from storage.
* Combines tool outputs and insights.
* Generates a coherent, natural response.

This makes the terminal feel less like code and more like **a conversation with an expert assistant**.

***

### **Why aigentOS is Different**

Unlike generalized AI frameworks, aigentOS is:

* **Purpose-Built**: Designed specifically for onchain AI agents.
* **Lightweight**: Fast, efficient, and doesn’t overload resources.
* **Accessible**: Anyone can use it, no technical skills required.
* **Scalable**: Built to expand as new MCP tools and integrations are added.

***

### **The Foundation of the Terminal**

With aigentOS, aigent.run delivers a **single, unified framework** for running agents.\
Every command, every interaction, every workflow in the terminal is powered by aigentOS, ensuring speed, security, and adaptability for all users.