· bingx integration

AlgoVault has 91.8%+ PFE Win Rate across 388,981+ signal calls, each Merkle-anchored on Base L2 (verifiable at ).

AlgoVault × BingX — Build Verifiable AI Trading Agents

AlgoVault MCP gives your agent a composite verdict in one call — direction, confidence, regime, and cross-venue funding/sentiment context — backed by a publicly auditable record anchored to Base L2. Pair it with BingX’s VST demo-trading environment and your agent has both the analytics brain and the execution venue.

Provenance: BingX publishes no official client SDK — its GitHub org (BingX-API) contains documentation repos and an AI-skills bundle, not a client library, and no BingX-published package exists on npm. The demo below therefore uses plain fetch + Node’s built-in crypto, with zero third-party dependencies. Demo execution runs against BingX’s VST (Virtual Simulated Trading) host https://open-api-vst.bingx.com, documented in BingX’s own org repo as “Production Simulated (Testnet) … use prod-vst for paper trading / testing without real funds”, with BINGX_TESTNET=true + MAINNET_BLOCKED=true wrapper guards. Verified 2026-07-21.

What you’ll build (90s read)

A runnable Node.js agent that:

  1. Calls AlgoVault MCP for a composite verdict on a chosen asset + timeframe.
  2. Reads the verdict’s call, confidence, regime, and reasoning fields.
  3. Hands the verdict to BingX’s VST demo environment, which validates a small order via the dedicated dry-run endpoint when the agent’s pre-configured policy fires.

The whole loop runs against BingX’s VST demo host — zero real-money risk in any code path.

Prerequisites (4 items)

  1. Node.js ≥ 22 (node --version to check).
  2. AlgoVault skills plugin installed:
    claude plugin install AlgoVaultLabs/algovault-skills
  3. BingX account (free signup at https://bingx.com). API key + secret from the API management console. Keys are read-only by default — enable trade permission explicitly, and IP-allowlist them.
  4. No SDK to install. The demo runs read-only out of the box with no keys at all; supplying keys unlocks the authenticated dry-run step.

Demo: Confidence-filtered entry validator on BTC (≤80 lines)

// examples/bingx/demo.mjs (excerpt — see file for full source)
import crypto from 'node:crypto';
import { getAlgoVaultVerdict } from '../_shared/algovault-helper.mjs';

const MAINNET_BLOCKED = true;
const BINGX_VST_BASE = 'https://open-api-vst.bingx.com';

if (process.env.BINGX_TESTNET !== 'true') {
  throw new Error('BINGX_TESTNET=true required — demo refuses to run against mainnet.');
}

const verdict = await getAlgoVaultVerdict({ coin: 'BTC', timeframe: '1h' });

if (verdict.call === 'BUY' && verdict.confidence > 70) {
  // Dry-run: validates signature + params, places nothing.
  await signedPost('/openApi/swap/v2/trade/order/test', {
    symbol: 'BTC-USDT', side: 'BUY', positionSide: 'LONG',
    type: 'MARKET', quantity: 0.001,
  });
}

Run it:

BINGX_TESTNET=true node examples/bingx/demo.mjs

Expected output (last 3 lines):

[bingx vst | contracts HTTP 200 | BTC-USDT tradable on the demo host]
[policy | fires=false (call===BUY && confidence>70)]
=== NO REAL ORDERS PLACED ===

Walkthrough (line-by-line — neutral narration)

The script does three things in order:

  1. Fetch the AlgoVault verdict. The helper opens an MCP session against https://api.algovault.com/mcp, calls get_trade_call with {coin, timeframe}, and returns the parsed call / confidence / regime / reasoning payload. Free tier covers up to 20 calls/day per IP — plenty for development.

  2. Apply the agent’s policy. When the verdict satisfies a pre-configured policy (e.g. the agent’s policy fires when confidence > 70 AND call === 'BUY'), the script proceeds to the execution branch. The policy lives entirely in your code — AlgoVault returns the analytics; the agent decides.

  3. Hand the order to BingX. The signed request goes to POST /openApi/swap/v2/trade/order/test, which validates the signature and parameters without placing a trade. The script aborts immediately if BINGX_TESTNET is not set — a hard guard against accidentally running against mainnet.

3 Recipes

Recipe 1 — Regime-gated entries

Call get_market_regime before acting. When the regime is RANGING, a trend-following agent can skip the entry rather than paying the spread on a chop signal. The regime arrives in the same composite response — no second provider, no extra call.

Recipe 2 — Validate before you place

BingX exposes POST /openApi/swap/v2/trade/order/test alongside the real POST /openApi/swap/v2/trade/order. Route every order through the test endpoint first and only submit on a clean response. It is the cheapest possible guard against a malformed signature or a bad positionSide, and it costs nothing.

Recipe 3 — Top up demo funds from code

The VST balance is adjustable via the API itself: POST /openApi/swap/v2/trade/getVst takes adjustType (0 increase / 1 decrease) and amount, rate-limited to 5/s per UID. A long-running backtest harness can therefore reset its own demo balance between runs without touching the web console.

⚠️ Production setup (real-money)

The demo above runs on the VST demo host only. Real-money setup requires:

  • KYC completion on BingX.
  • Production API keys with only the permissions your agent needs (no withdrawals, IP-allowlisted). Keys are read-only until you explicitly grant trade permission.
  • Risk controls: per-order size cap, daily loss limit, kill switch.
  • Position monitoring: a separate agent or watchdog that tracks open positions independently.
  • Awareness of the order rate limit — 10/s per UID and 3/s per IP on the place-order endpoint.

See examples/bingx/README.md for the full real-money checklist. AlgoVault provides analytics; your agent and your risk policy decide what (if anything) to execute.

Why AlgoVault?

  • Composite verdict, not raw indicators. One JSON response replaces 26-indicator vote-counting.
  • Cross-venue intelligence. Funding spreads, regime, and sentiment fused across 12 exchanges — not derivable from any single-venue API.
  • Publicly verified. Every signal anchored to Base L2 via Merkle proof. Verify before you subscribe.

91.8% PFE Win Rate · 388,981+ calls · 102+ on-chain batches → view live track record

Install

claude plugin install AlgoVaultLabs/algovault-skills

Once installed, every Skill in the pack is one-line invokable from Claude Code, Cowork, or any MCP-compatible client.

Tutorial © AlgoVault Labs · MIT licensed · Provenance verified 2026-07-21 · BingX publishes no official client SDK; demo uses plain fetch + node:crypto