> For the complete documentation index, see [llms.txt](https://olaxbt-docs.gitbook.io/olaxbt-doc/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://olaxbt-docs.gitbook.io/olaxbt-doc/project-info/olaxbt-agent.md).

# OLAXBT Agent

## Vision

OLAXBT Agent is not another trading bot — it is the first adaptive, user-owned intelligence layer for digital assets. By combining continuous reinforcement learning with on-chain and off-chain context, each agent evolves in real time to match its owner’s unique trading style, risk tolerance, and market philosophy.

## Core Concept

The Living Co-PilotUnlike static indicators or rule-based bots, an OLAXBT Agent is a persistent, stateful entity that:

* Learns from every trade its owner executes (win or loss)
* Ingests live on-chain flows, order-book dynamics, and sentiment signals
* Remembers past conversations and user feedback
* Autonomously refines its decision framework without ever taking custody of funds

The result is a co-pilot that gets sharper the longer you use it, effectively turning personal trading experience into compounding intellectual capital.

## Powered by Model Context Protocol (MCP)

MCP is the open framework that allows agents to be modular, composable, and marketable:

* Agents are built from interchangeable “skills” (technical analysis, funding-rate arbitrage, KOL sentiment tracking, narrative detection, etc.)
* Users can assemble agents without writing code using the no-code IDE
* Finished agents can be listed on the MCP Marketplace and monetized as Agent-as-a-Service (AaaS)
* Third-party developers earn revenue when their skills or full agents are used

This transforms trading tools from closed products into an open, collaborative economy.