Learn

Connect any LLM to your WICK account

5 min read · Last verified June 21, 2026

What this actually is

WICK FUNDED is a 100% simulated trading environment. No real funds are exposed to markets. The API and MCP endpoint described here connect to that simulation — not a live brokerage, not your bank account.

With that clear: WICK exposes a documented REST API and an MCP endpoint so you can wire your own language model to your own simulated account. WICK supplies the account infrastructure, the simulator, the drawdown enforcement, and the toolset. You supply the strategy and the brain. The platform is deliberately unopinionated about which model you use or how you structure your logic.

This is a power feature aimed at traders who already build with LLMs. It is not a bot marketplace, not a copy-trading product, and not a shortcut around doing the evaluation work.


Getting a key

Every API call is authenticated with a Bearer token in the format wk_live_.... You generate this from your WICK dashboard under API Access.

One key, one account. The key is scoped to your account only — you cannot use it to read or trade on anyone else's account, and sharing it is explicitly prohibited (someone else using your key means someone else trading your funded account, which violates the terms under which payouts are calculated).

Keep the key out of version control. Store it in an environment variable or your secret manager of choice.


The MCP endpoint

If you use an MCP-compatible client — Claude Desktop, Cursor, or a custom agent built on the MCP protocol — the setup is a single config block:

{
  "mcpServers": {
    "wick": {
      "url": "https://wickfunded.com/api/mcp",
      "headers": {
        "Authorization": "Bearer wk_live_your_key_here"
      }
    }
  }
}

Add that to your client's MCP config file, restart, and your agent now sees the WICK tools alongside whatever else you have wired up.


What the tools give you

Once connected, your agent can call the following categories of tools:

CategoryWhat you can do
AccountRead balance, equity, drawdown levels, track status
PositionsList open positions, individual position detail
JournalRead trade history, P&L breakdown, timestamps
CandlesFetch OHLCV data for any supported symbol and timeframe
OrdersOpen, close, and modify trades

The read tools are useful for building agents that reason about their own state — checking current drawdown before sizing a new position, reviewing the journal to avoid repeating a losing pattern, or pulling candles to compute an indicator inline.

The write tools (open/close/modify) are what make autonomous trading loops possible. A typical pattern: agent reads current equity and open positions, checks candles for a signal condition, decides whether to open a trade, and fires the order — all in a single LLM turn or a short agentic loop.


Practical wiring patterns

Claude Desktop / Claude Code: Add the MCP block above to ~/.claude/claude_desktop_config.json. Your Claude instance now has WICK as a tool server. You can prompt it: "Check my current drawdown and tell me whether I have room to open a 1% risk trade on EURUSD" and it will call the account and candles tools before answering.

Cursor: Same MCP config pattern. Useful if you are writing strategy code and want the agent to run live tests against your simulated account during development.

Custom agent: If you are building a Python or TypeScript agent, the REST API is the direct path. Every tool exposed via MCP has a corresponding REST endpoint. Authenticate with Authorization: Bearer wk_live_..., send JSON, get JSON back. The API docs in your dashboard list full request/response schemas.

ChatGPT (custom GPT or function calling): You can define the WICK endpoints as custom functions in a GPT tool schema. Slightly more setup than the MCP path, but the same underlying API.


Drawdown mechanics your agent must respect

Your agent is operating inside a live evaluation or funded account with real drawdown rules. Getting these wrong ends the account — understand them before automating.

The API exposes current drawdown levels in the account endpoint. Any well-built agent should read this before opening a position, not after.


What is not allowed

The rules are straightforward:


Honest expectations

The API does not make strategy generation easier. It removes the friction of manual order entry and makes it practical to run a systematic approach at higher frequency or with tighter rules-based logic. Whether the strategy itself is sound is still entirely on you.

Payouts on WICK FUNDED accounts start at 80% and increase by 5% per completed paid payout cycle up to 95%. Those payouts are promotional rebates from operator treasury based on simulated performance — not withdrawals from a trading account. KYC is required at first payout only, minimum is $50, and the delivery mechanism is USDT TRC20.

Nothing in this article is financial advice.

If you are building something serious with the API, start with read-only tools, validate that your agent is interpreting account state correctly, then enable order execution. The simulator gives you a genuinely isolated environment to do that safely before any payout thresholds are in play.