crypto-quant-signal-mcp

Quick start (30 seconds)

No code. No install. No API key to start — sign in free when you need more. The server speaks Streamable HTTP at https://api.algovault.com/mcp — any Model Context Protocol client connects directly.

1. Add the connector. In Claude → Customize → Connectors → + → Add custom connector:

If Claude asks how to sign in, choose Sign in when needed. When Claude asks, click Connect and sign in with your email.

2. Ask for a call. In plain language:

"Get me a trade call for ETH on the 4h timeframe"

Your Claude now has a quant analyst built in. Prefer local? Run npx -y crypto-quant-signal-mcp.

Running locally on npm 12+? npm v12 disables dependency install scripts by default. AlgoVault's optional local SQLite mode uses the native better-sqlite3 module — if you install it into a project, run npm approve-scripts (or npm install --allow-scripts) to build it. The hosted HTTP transport (api.algovault.com/mcp) needs no local build.

Related MCP server: Web3 Signals — Crypto Signal Intelligence

What one call returns

One API call. One verdict. Not 8 raw indicators. get_trade_call returns a directional BUY / SELL / HOLD with a confidence score and the detected market regime — a composite verdict, not a data dump for your agent to interpret.

Under the hood, a self-tuning model fuses momentum, trend structure, derivatives positioning, open interest, and volume into one weighted call. Every call reports market regime — trending, ranging, or volatile — alongside direction and confidence. Directional calls fire only when conviction clears the threshold; roughly 99% of evaluations return HOLD.

{
  "call": "BUY",
  "confidence": 78,
  "price": 84250.50,
  "indicators": {
    "funding_rate": 0.0001,
    "funding_24h_avg": 0.00008,
    "funding_state": "NORMAL",
    "oi_change_pct": 2.4,
    "volume_24h": 2381602633,
    "trend_persistence": "HIGH",
    "breakout_pending": "INACTIVE"
  },
  "regime": "TRENDING_UP",
  "reasoning": "Trending regime, upward bias. Funding pressure mild. Volatility neither expanding nor compressed. Trend persistence elevated; momentum structure. Strong conviction from aligned signals.",
  "timestamp": 1712764800,
  "coin": "BTC",
  "timeframe": "1h",
  "also_see": [
    { "coin": "ETH", "timeframe": "1h", "confidence": 82 },
    { "coin": "SOL", "timeframe": "15m", "confidence": 73 }
  ],
  "_algovault": {
    "version": "1.30.0",
    "tool": "get_trade_call",
    "compatible_with": []
  }
}

The _algovault block is the envelope every tool response carries: the server version, the tool that answered, your quota, and compatible_with — the companion packages that accept this object directly. It projects from a machine-verified registry, so it lists a package only once that package is published and runnable. It is [] today, and that is the truthful answer: there is no companion package yet.

Live, verifiable track record

The full record is public and on-chain — no cherry-picking, no survivorship bias:

• Live dashboard — algovault.com/track-record

• Merkle batch verifier — algovault.com/verify

• Anchor contract on Base L2 — 0x6485…0f81

• All batches (raw JSON) — api.algovault.com/api/merkle-batches

AlgoVault is also a verified agent on the ERC-8004 Identity Registry (Base L2), agentId 44544 — a portable, on-chain handle AI orchestrators can resolve.

Works with your stack

AlgoVault is drop-in for every MCP-spec client, every major agent framework, and every official exchange Agent Trade Kit — no SDK, no wrapper. It serves Streamable HTTP at https://api.algovault.com/mcp; tools/list + resources/list is the API surface.

MCP clients.

Agent plugin hosts. The npm package ships plugin.json, mcp.json and .cursor-plugin/plugin.json, so an Agent-Plugins-compatible host — Cursor included — can read the server definition out of the installed package instead of being handed a config block to paste.

Agent frameworks. First-party tutorials pair AlgoVault with each framework's canonical MCP adapter — copy-pasteable demo code, no SDK.

Each demo is runnable as python examples/<framework>/demo.py BTC 4h — gets a real BUY/SELL/HOLD verdict from api.algovault.com/mcp, prints it. ≤5 minutes to first call.

Pre-trade gate. Make pre_trade_check a step your agent cannot skip. In Claude Code and Codex, the AlgoVault pre-trade gate blocks an order until the check has run for that coin and side, and stops it on WAIT. The LangChain and Freqtrade recipes run the check before an order and block it when the check cannot run.

Exchange Agent Trade Kits. AlgoVault returns the analytics; your agent's risk policy decides what to execute. All demos run testnet/demo only. For Binance, the Agent OS path is the recommended default for new integrations — OAuth at connect time, no API keys on the machine; the key-based kit remains right for custom order types.

Tools & resources

The MCP tools live at https://api.algovault.com/mcp. Every asset works across the full supported timeframe range, on major crypto perpetual venues.

• get_trade_call (alias get_trade_signal) — composite BUY/SELL/HOLD verdict with confidence + regime, any asset, any timeframe.

• scan_trade_calls — scans the top-N perps by open interest on a venue; returns every actionable call in one shot.

• scan_market_regimes — screens a whole venue by market regime in one call: every coin whose regime matches your filter, with confidence and how long the regime has held. One call per returned row (at least one per request); coins that don't match are free.

• pre_trade_check — checks an order against your own rules before you place it: each rule passes or fails with its numbers, then GO, CAUTION or WAIT, with isolated liquidation per venue against your stop. Read-only: it places no orders.

• scan_funding_arb — cross-venue funding-rate spreads across 7 venues (Hyperliquid, Binance, Bybit, Gate, KuCoin, Aster, OKX), ranked. The only MCP server doing multi-exchange derivatives arbitrage.

• get_market_regime — classifies TRENDING_UP / TRENDING_DOWN / RANGING / VOLATILE for strategy selection.

• search_knowledge + chat_knowledge — free BM25 search and grounded Q&A over the full knowledge bundle.

• get_track_record — the published track record as a tool call, for clients that bridge tools but not resources. include opens the per-asset, per-venue and recent-call sections.

Performance is exposed as a read-only MCP resource: performance://signal-performance (aggregated PFE win rate, never raw outcomes). Full parameter reference at algovault.com/docs.

Every answered pre_trade_check carries a receipt you can verify through the read-only resource verify://check/{id} or GET /api/verify-check?id=, against its root anchored on Base.

Skills (20 ready-to-use Anthropic Agent Skills)

Single-prompt wrappers over 1–3 tool calls — regime gating, multi-timeframe consensus, funding-arb monitoring, and more. Install with claude plugin install AlgoVaultLabs/algovault-skills; browse at algovault.com/skills.

Pricing

Quota-only tiers. Every tier gets all venues, all assets, all timeframes — you pay for call volume, nothing else.

Two meters per paid tier, monthly and daily, enforced independently: a call is refused when either is exhausted, and the daily window is a UTC calendar day resetting at 00:00 UTC. Quota is counted per call, regardless of verdict.

Subscriptions: Sign up at api.algovault.com/signup. Starter unlocks 10,000 calls/mo (up to 1,000/day) for $9.99/mo, or $39.90 for six months prepaid — add ?interval=6month to pick the prepay term. API key delivered instantly after checkout.

Pro 6-month is currently $129 — limited-time pricing; subscribe now and renewals keep your price.

x402 micropayments: AI agents pay per HTTP call with USDC on Base — no signup, no API key, no billing. The payment receipt is the credential. See x402.org.

What's new in v1.34.0

• 📌 Save your rules once. Save your pre_trade_check rules with PUT /api/pre-trade-rules and your API key, and every check that key makes applies them. Rules you send with a call still win. A rule from your saved profile shows source: "saved", and saved_rules says whether a profile applied.

• 🚧 Make the check a step your agent cannot skip. In Claude Code and Codex, the AlgoVault pre-trade gate blocks an order tool call until pre_trade_check has recently checked that coin and side, and blocks it when the check returns WAIT. In LangChain and Freqtrade, a recipe runs the check before the order.

• 🧾 A receipt with every answered check. An answered pre_trade_check carries a ledger receipt. Once its root is anchored on Base, verify://check/{id} or GET /api/verify-check?id= proves the verdict your agent got was recorded unchanged.

Upgrading from v1.33.x — additive. No tool or tool parameter was added, renamed or removed. pre_trade_check responses gain saved_rules and ledger, and a rule can now show source: "saved". New: the resource verify://check/{id} and the routes /api/pre-trade-rules and GET /api/verify-check. Market regime direction follows a corrected rule, so some regimes read differently and regime ages in scan_market_regimes restarted.

v1.33.x highlights (recap)

• 🛡️ Check an order against your own rules before you place it. pre_trade_check takes the order you intend to place and your limits (maximum leverage, loss at your stop, accuracy needed to break even, funding) and returns each rule's pass or fail with its numbers, then GO, CAUTION or WAIT. It also shows where isolated liquidation sits on each venue that publishes its margin tiers, and whether it hits before your stop.

• 🔐 Sign in from inside Claude. Add AlgoVault as a custom connector and choose Sign in when needed. When Claude asks, click Connect and sign in with your email. On a paid plan, sign in with your subscription email and your plan applies.

Upgrading from v1.32.x — one tool was added, pre_trade_check; no tool or parameter was renamed or removed. A call with an argument the tool doesn't take is now refused, unbilled, and the refusal names the valid parameters; the published schemas never allowed extra arguments. Added AlgoVault in Claude before? Remove it and add it again, choosing Sign in when needed.

v1.32.x highlights (recap)

• 🧭 Screen a whole venue by market regime in one call. scan_market_regimes returns every coin whose regime matches your filter — TRENDING_UP, TRENDING_DOWN, RANGING or VOLATILE — with confidence, trend strength, ADX, volatility ratio and how long the regime has held. Filter by regime, confidence, asset class, liquidity, regime age or your own watchlist, then sort by confidence or by the newest regime change. You pay one call per returned row (at least one per request); coins that don't match are free. A venue or timeframe it can't serve right now is refused with the alternatives, and isn't billed. The screener is served from the hosted endpoint; a local install lists it and refuses the call with a pointer there.

Upgrading from v1.31.x — additive. One tool was added, scan_market_regimes; no tool or parameter was renamed or removed. GET /api/plans/public gains two blocks under free: keyless (the allowance without a key) and telegram (the Telegram bot's own allowance, counted in alerts). One free-tier allowance changed: 20 calls a month without a key; a free key gives 200 (up to 100/day).

Refresh your MCP client to pick up this release. MCP clients cache tools/list at session start — Claude.ai/Desktop: toggle the connector off+on; Cursor/Cline: restart the MCP server connection.

Privacy

Local mode: zero telemetry — call history stays on your machine. Remote mode: request metadata logged for analytics (IP hashed, never stored raw). See the privacy policy.

License

MIT

Disclaimer: AlgoVault provides directional entry interpretation for AI agents. Exit timing is determined by your agent or strategy. This is not financial advice. Past performance does not guarantee future results.