TL;DR
- Kraken CLI is an open-source, single-binary execution engine built by Kraken that gives AI agents and developers direct, native access to crypto markets — supporting spot trading, futures, staking, subaccount transfers, and WebSocket streaming across 134 commands, all via a zero-dependency Rust binary with clean NDJSON output designed for machine consumption.
- Kraken CLI includes a built-in MCP (Model Context Protocol) server, making it the first crypto exchange CLI natively compatible with agentic coding tools like Claude Code, Codex, and Cursor, as well as terminal-native agent environments like OpenCode and OpenClaw — enabling AI agents to instantly understand and execute exchange operations without custom API wrappers, nonce management, or HMAC signing.
- Kraken CLI ships with a local paper trading engine that lets AI agents safely test trading logic against live market data with no financial risk — running a local state engine that tracks simulated balances, executes limit and market orders against the live ticker, and calculates unrealized P&L entirely offline.
Built for the way AI agents actually work
To accelerate this transition to an agentic future, we are thrilled to open-source the Kraken CLI: a powerful, single-binary execution engine purpose-built to give AI models (and developers) safe, native, and instant access to crypto markets.
Whether you’re an algorithmic trader, a terminal power user, or a developer building LLM-powered financial tools, the Kraken CLI bridges the gap between raw AI intelligence and live market execution.
Why we built it: moving beyond raw APIs
APIs are designed for systems. But when you ask an AI agent to interact with a raw crypto exchange API, it struggles. It has to write custom Python scripts, manage cryptographic nonces, handle complex HMAC-SHA512 payload signing, and build its own retry logic for rate limits.
This process is highly error-prone and burns through valuable context window tokens.
We built the Kraken CLI to abstract away the boilerplate. Written entirely in Rust, it is a zero-dependency binary that handles the infrastructure of exchange connectivity for you, allowing you (and your agents) to focus purely on trading logic.
Key features for developers
Here’s what’s under the hood:
Built-in MCP server
The CLI includes native support for the Model Context Protocol (MCP). By simply running kraken mcp, the CLI transforms into a secure, self-describing plugin for agentic coding tools like Claude Code, Codex, and Cursor, as well as terminal-native environments like OpenCode and OpenClaw. Your AI agent instantly understands the schemas for all 134 available commands.
Local paper trading engine
Safely test your agent’s trading logic against live market data without risking a cent. The Kraken paper suite runs a local state engine that tracks simulated balances, executes limit/market orders against the live public ticker, and calculates unrealized P&L offline.
Universal shell compatibility
If your agent doesn’t support MCP, it can simply use the terminal. The CLI guarantees clean, machine-readable NDJSON output (-o json) across all commands, making it incredibly easy to pipe data into jq or feed it directly back into an LLM’s context window.
Full exchange surface area
This isn’t just a basic price ticker. The CLI supports spot trading, futures trading, staking/earn allocations, subaccount transfers, and even WebSocket streaming.
Enterprise-grade resilience
Built-in spot counter/decay and futures token-bucket rate limiting ensure your automated workflows don’t crash when hitting exchange limits.
One-line installation
curl --proto '=https' --tlsv1.2 -LsSf https://github.com/krakenfx/kraken-cli/releases/latest/download/kraken-cli-installer.sh | sh
The future is agentic
The Kraken CLI is where the industry is heading: a world where AI agents are first-class participants in financial markets, not an afterthought bolted onto legacy infrastructure. By open-sourcing this tool, we’re inviting developers, algorithmic traders, and AI builders to help shape what that future looks like – and to start building it today.
Check out the full documentation, source code, and agent prompt examples on GitHub:
Trading derivatives and other financial instruments, including leveraged financial instruments, involves significant risks and is not appropriate for all investors. See our Risk Disclosure to learn more. Geographic restrictions apply.