← Back to directory
O

OKX Agent Trade Kit

Community
Connect AI assistants directly to your OKX account via MCP, with local, secure trading.
Category
Other #20 of 230
Stars
★ 110 Popular
Transport
stdio (local process)
Runtime
Node.js 18+
Credentials
Optional API key
License
MIT
Last commit
68FMRS · C

OKX Agent Trade Kit is a feature-rich, open-source, privacy-focused tool for OKX trading, offering a wide range of tools and CLI support, ideal for AI-assisted trading. Safety controls (read-only mode, rate limiting) enhance security, but trading risks remain the user's responsibility.

Strongest · Documentation 16/20 Weakest · Reliability 10/20

Reliability
10/20
Security and permissions
14/20
Maintenance
15/20
Documentation
16/20
Setup experience
13/20
Why each score
Reliability 10/20
Source structure is reasonable, with core, MCP, and CLI packages, CI workflow and test files present, but not executed. Tools claimed as 145 but no full list provided. Structured error handling exists, indicating some design for errors. However, actual runtime stability cannot be confirmed, so deduction.
Security and permissions 14/20
API keys stored locally, supports read-only mode, rate limiter and module filtering, but no explicit data-flow disclosure or confirmation for dangerous operations found. No malware or real secrets in install examples. Main risks visible, but permission and confirmation are incomplete.
Maintenance 15/20
Active project with CI badges, codecov, npm releases, changelog and contributing guide, clear MIT license. However, version governance details and dependency update strategies not fully shown, leaving room for improvement.
Documentation 16/20
Documentation is comprehensive, including quickstart, module docs, CLI reference, configuration guide, FAQ, security policy. Bilingual. But some limitations (like rate limit details) and cost info not fully disclosed across all docs.
Setup experience 13/20
Clear install guide, supports multiple clients (Claude, Cursor, etc.), interactive config wizard and one-line install script. Requires Node.js >=18, but lacks detailed platform compatibility. No verified execution evidence, so cap at 15.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

  • Traders who want to leverage AI assistants for cryptocurrency trading
  • Users who value local operation and data privacy
  • Developers who want to integrate trading functionality into MCP-compatible clients like Claude Desktop or Cursor
  • Users who need automated trading strategies (grid, DCA)

Not for

  • Beginners unfamiliar with cryptocurrency trading (requires basic knowledge)
  • Users who need cloud or remote server operation (designed as a local process)
  • Users who need support for other exchanges (OKX only)
  • Users who prefer a GUI over CLI or AI interaction

Required permissions

  • Read market data (no authentication required)
  • Trading operations (requires OKX API credentials): spot, swap, futures, options, etc. - place, cancel, query
  • Account management: query balance, bills, positions, fee rates, etc.
  • Earn operations: purchase, redeem, staking, etc.
  • Trading bot operations: create and manage grid and DCA strategies

Risks and side effects

  • API key leakage could lead to asset loss; store keys securely
  • Trading operations carry risk; AI may execute unintended commands; consider using read-only mode initially
  • Local process depends on Node.js environment; keep npm packages updated
  • Market volatility can lead to losses; use caution

Setup

Before you start

Runtime:Node.js 18+

  1. Install Node.js >= 18.
  2. Run npm install -g @okx_ai/okx-trade-mcp @okx_ai/okx-trade-cli.
  3. Run okx config init to configure OKX API credentials (interactive wizard).
  4. Register the MCP server with your AI client using okx-trade-mcp setup, e.g., okx-trade-mcp setup --client claude-desktop.

Check that it works

After running okx-trade-mcp setup --client claude-desktop and restarting the client, module tools such as spot_place_order or account tools should appear in the tool list; starting with --modules market verifies the connection without API keys.

Troubleshooting

  1. Ensure Node.js >= 18 is installed
  2. Check API key configuration is correct (`okx config init`)
  3. Inspect the structured error output (for tool calls or CLI), which includes error code and endpoint
  4. Use correct tool naming: MCP tools use underscores (e.g., `swap_place_order`), CLI subcommands use spaces (e.g., `okx swap place`)

Things to try

Once connected, you can ask your AI assistant things like:

  • Get the latest ticker and funding rate for BTC-USDT
  • Show my OKX account balance and open positions
  • Place a market buy order for 100 USDT of BTC-USDT
  • Run okx-trade-mcp setup --client claude-desktop for my AI client

Use cases

Use AI assistants for OKX trading operations, such as market data queries, order placement, and position management
Automate trading strategies like grid or DCA using trading bot tools
Analyze market trends with Smart Money signals
Execute trading commands quickly from the terminal using the CLI

Supported clients

Claude Desktop
Cursor
Claude Code
VS Code

Listed from the project's documentation, not tested by this site.

Overview

OKX Agent Trade Kit is an open-source toolkit that connects AI assistants to your OKX trading account via the Model Context Protocol (MCP). It provides 145 tools across 11 modules covering market data, spot, swap, futures, options, account, earn, trading bots, news, and smart money signals. It runs as a local process with API keys stored only on your machine; no cloud services, no data leaving your device. Includes both an MCP server and a CLI, with safety controls like --read-only flag, per-module filtering, and built-in rate limiter.

Similar servers

BioMCP 75 · B

One binary. One grammar. Evidence from the biomedical sources you already trust.

★ 654 · Tools 11 Compare with this →

Source revision f619ddf3bb4a Data synced 2026-10-12 Read the FMRS scoring method