If you're still manually refreshing a dashboard to check a funding rate or clicking through three sub-menus to adjust a stop-loss, you're basically trading via carrier pigeon.
The industry has moved past using AI as a mere search bar. Today, over 42% of financial professionals use agentic AI to manage complex workflows, and 84% agree that open-source models are the only logical path forward. The market is moving away from manual monitoring, and we think it's time your trading interface did the same.
Today, we are launching the Toobit AI Agent Trade Kit, an open-source framework designed to turn your favorite LLM into a functional market participant.
Conversations are the new terminal
We built this kit because dashboards are becoming a legacy habit. By utilizing the Model Context Protocol (MCP), we've developed a series of modular files that teach AI agents how to handle specific tasks without additional programming. This gives agents like Claude and ChatGPT the "hands" they need to handle the heavy lifting of your Toobit workflow:
-
Market data: Instead of you hunting for price history and order book depth, your agent simply watches the tape for you. It pulls the public data it needs to stay informed, and it doesn't even need a login to do it.
-
Trading operations: This is where the agent earns its keep. You define the strategy, and the agent handles the actual clicking. From opening a spot position to managing risk on a high-leverage perpetual, it bridges the gap between a "plan" and a "trade."
-
Assets portfolio: Checking your balance across multiple accounts shouldn't feel like a chore. You can ask your agent for a real-time P&L or a transaction history update in plain English, and it will crunch the numbers while you focus on the next move.
Your keys, your agent, your computer
We know the standard crypto anxiety: "If I give an AI my keys, where do they go?"
The answer is nowhere. We designed this kit to run entirely locally. Your credentials live in a local config file on your machine, and transaction signing happens right there on your hardware. If you don't trust the AI with your full balance yet, you can toggle specific modules or keep the agent in read-only mode.
We believe in delegated intelligence, not surrendered control. You set the boundaries; the agent does the work.
Open source or bust
The future of finance isn't going to be locked behind a proprietary black box. That's why the AI Agent Trade Kit is distributed under the MIT License and is live on GitHub right now.
Whether you want to use a chat interface to manage your morning positions or run a CLI for automated terminal management, the infrastructure is ready.
We are moving into an era where the best traders aren't the ones with the fastest clicking fingers, but the ones with the best-equipped agents. It's time to step away from the screen and let your strategy move forward on its own.
Ready to build your digital teammate? Check out the AI Agent Trade Kit page to get started.
