Developer API & LLM integration

Bring your own model. We bring the simulator.

WICK FUNDED exposes a documented REST API and a streaming MCP endpoint so a trader can wire their own LLM agent — Claude, ChatGPT, Gemini, a custom model, anything — to their own simulated account. We supply the tools; you supply the strategy and the brain. Same simulator, same rules, same execution engine as the web UI.

Full policy: Terms §8 — Personal Automation · /rules for breach detection details.

Quickstart

1. 1. Go to Dashboard → Developer → API keys and create a token. It looks like wk_live_… and is shown once.
2. 2. Pick a scope: read for data only, execute to place / close / modify trades.
3. 3. Authenticate every request with a Bearer header. That's the entire auth model.
4. 4. Drop the OpenAPI URL (/api/v1/openapi) into Postman, Cursor, or Claude and you have full tool definitions for free.

curl — open a position

curl -X POST https://wickfunded.com/api/v1/trade/open \
  -H "Authorization: Bearer wk_live_xxxxxxxxxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "account_id": "your-account-uuid",
    "symbol":     "EURUSD",
    "side":       "buy",
    "lot_size":   0.10,
    "sl_price":   1.0820,
    "tp_price":   1.0900
  }'

Python — full read-then-execute loop

import os, requests

WK = os.environ["WICK_API_KEY"]
BASE = "https://wickfunded.com/api/v1"
H = {"Authorization": f"Bearer {WK}"}

# 1. find the active funded account
acct = requests.get(f"{BASE}/account", headers=H).json()["accounts"][0]

# 2. pull 200 candles of EURUSD M5
candles = requests.get(
    f"{BASE}/candles",
    headers=H,
    params={"symbol": "EURUSD", "tf": "M5", "limit": 200},
).json()

# 3. (your strategy here)
# decision = my_llm_or_strategy(candles)

# 4. execute
r = requests.post(
    f"{BASE}/trade/open",
    headers=H,
    json={
        "account_id": acct["id"],
        "symbol":     "EURUSD",
        "side":       "buy",
        "lot_size":   0.10,
        "sl_price":   1.0820,
        "tp_price":   1.0900,
    },
)
print(r.status_code, r.json())

TypeScript — expose WICK as an LLM tool

Same pattern works with Anthropic tool_use, OpenAI tools, or any tool-calling model.

import OpenAI from "openai";          // or @anthropic-ai/sdk
const openai = new OpenAI();

const WK   = process.env.WICK_API_KEY!;
const BASE = "https://wickfunded.com/api/v1";

// Expose WICK actions as OpenAI tools — same pattern works with Claude tool_use
const tools = [{
  type: "function" as const,
  function: {
    name: "wick_trade_open",
    description: "Open a market position on the user's WICK FUNDED simulator",
    parameters: {
      type: "object",
      required: ["symbol", "side", "lot_size"],
      properties: {
        symbol:   { type: "string", example: "EURUSD" },
        side:     { type: "string", enum: ["buy", "sell"] },
        lot_size: { type: "number", minimum: 0.01, maximum: 50 },
        sl_price: { type: "number" },
        tp_price: { type: "number" },
      },
    },
  },
}];

const chat = await openai.chat.completions.create({
  model: "gpt-4o",
  tools,
  messages: [
    { role: "system", content: "You are a discretionary trader. Use the WICK simulator tool when ready." },
    { role: "user",   content: "EURUSD is bouncing off the 1H demand zone — go long 0.10 lot, SL 1.0820, TP 1.0900." },
  ],
});

// When the model calls the tool, forward the args to WICK
for (const call of chat.choices[0].message.tool_calls ?? []) {
  const args = JSON.parse(call.function.arguments);
  const account_id = "your-account-uuid";
  const res = await fetch(`${BASE}/trade/open`, {
    method: "POST",
    headers: { Authorization: `Bearer ${WK}`, "Content-Type": "application/json" },
    body: JSON.stringify({ ...args, account_id }),
  });
  console.log(await res.json());
}

MCP — one-click Claude / Cursor integration

If your client speaks the Model Context Protocol, point it at https://wickfunded.com/api/mcp and the tools auto-register. Claude Desktop / claude.ai connect with a single OAuth click — no key to paste; you log into WICK and approve. Prefer an API key (Claude Code CLI, Cursor, scripts)? Use a Bearer token instead. Both below.

# ── Option A — One click (Claude Desktop / claude.ai) ──
# Settings → Connectors → Add custom connector. Paste ONLY the URL:
#
#     https://wickfunded.com/api/mcp
#
# Click "Connect" → you'll be sent to WICK, you log in, approve the scope
# (read, or read + execute), and you're done. No API key to paste, no JSON.
# Leave the OAuth Client ID / Secret fields BLANK — registration is automatic.

# ── Option B — API key (Claude Code CLI / Cursor / scripts) ──
# Create a key at Dashboard → Developer, then:
#
#     claude mcp add wick --transport http https://wickfunded.com/api/mcp \
#       --header "Authorization: Bearer wk_live_xxxxxxxxxxxx"
#
# or in a config file:
#   { "mcpServers": { "wick": {
#       "url": "https://wickfunded.com/api/mcp",
#       "headers": { "Authorization": "Bearer wk_live_xxxxxxxxxxxx" } } } }

# Tools exposed automatically: wick_account, wick_positions, wick_pending,
# wick_journal, wick_consistency, wick_symbols, wick_candles, wick_trade_open,
# wick_trade_close, wick_trade_pending, wick_trade_cancel, wick_trade_modify.
#
# Same rate limits + audit log as the REST API. Say things like:
#   "Check my WICK account and tell me my risk on the open EURUSD trade."
#   "Close half my GBPUSD position."

Endpoint reference

Live machine-readable spec: /api/v1/openapi.

Rate limits

Per API key, applied as a dual token bucket (per-second + per-hour). The same budget is shared whether you call the REST API or the MCP endpoint:

• 10 requests / second by default
• 100 requests / hour by default
• Every response carries X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset
• Burst over the cap → HTTP 429 with Retry-After
• Need more? Open a key with a higher cap from Settings (we approve larger budgets manually)

Audit log

Every call carrying the execute scope is permanently logged: timestamp, IP, key ID, request body, response status, latency, and (when supplied) the User-Agent or agent provider hint. You can browse your own log in Dashboard → Developer → API keys → Audit.

We do not use these logs to second-guess your strategy — they exist for anti-abuse, KYC/AML, and to give you a forensic timeline if a trade misbehaves.