SqueezeOS

For AI agents: Two live MCP servers with 44 tools. Pay per call in USDC on Base or RLUSD on XRPL via x402/HTTP-402.

Server	URL	Tools	Protocol
SqueezeOS MCP	https://squeezeos-api.onrender.com/mcp	33	MCP JSON-RPC 2.0
402Proof MCP	https://four02proof.onrender.com/mcp	11	MCP JSON-RPC 2.0

x402 flow: GET /api/{endpoint} → HTTP 402 + payment terms → pay USDC → retry with X-PAYMENT header → 200.
Free discovery: agents.json · agent.md · llms.txt

MCP Server · x402 payment-gated · RLUSD on XRPL · 33 tools · Live data only

SqueezeOS is an institutional-grade AI trading intelligence platform for autonomous agents. Premium endpoints are pay-per-call via 402Proof — agents pay RLUSD on the XRP Ledger and receive a 1-hour access token. No API keys. No subscriptions. No accounts.

Live MCP endpoint: https://squeezeos-api.onrender.com/mcp
Free demo: curl https://squeezeos-api.onrender.com/api/demo/council
Agent guide: https://squeezeos-api.onrender.com/llms.txt

Quick Start (30 seconds)

# 1. Hit free demo — see exact paid response format
curl https://squeezeos-api.onrender.com/api/demo/council

# 2. Connect as MCP server (Claude, GPT, any MCP client)

{
  "mcpServers": {
    "squeezeos": {
      "url": "https://squeezeos-api.onrender.com/mcp",
      "transport": "streamable-http"
    }
  }
}

Example Response

{
  "symbol": "IWM",
  "verdict": {
    "directive": "BUY (IGNITION)",
    "bias": "BULLISH",
    "confidence": 87,
    "regime": "ALPHA_EXPANSION",
    "thesis": "Gamma flip confirmed above $198. VPIN at 0.71 — institutional order flow dominant. SML Fractal Cascade locked: depth-3 anchors aligned. Options sweep detected: 4,200 contracts 200C, $1.2M premium. Battle Computer consensus: 6/7 engines bullish.",
    "targets": { "tp1": 201.50, "tp2": 204.00, "stop": 196.80 },
    "engines": {
      "gamma_flow": 92, "vpin": 88, "fractal_cascade": 91,
      "options_sweep": 85, "battle_computer": 86, "dark_pool": 79
    }
  },
  "data_sources": ["Tradier options chain", "Alpaca OHLCV", "XRPL on-chain"],
  "cached": false,
  "timestamp": "2026-06-05T14:32:11Z"
}

MCP Tools (33 total)

Free Tools

Paid Tools (RLUSD via x402)

Payment Flow (x402)

1. Call get_invoice(endpoint_id) → { pay_to, amount, memo_hex }
2. Send RLUSD on XRPL to pay_to with memo_hex as MemoData
3. Call verify_payment(invoice_id, tx_hash, agent_wallet) → access_token
4. Call any paid tool with payment_token: <access_token>
5. Token valid 1 hour. Reuse across all tools without re-paying.

Payment network: XRPL mainnet
Payment asset: RLUSD (issuer: rMxCKbEDwqr76QuheSUMdEGf4B9xJ8m5De)
Token TTL: 1 hour (HMAC-SHA256, wallet-bound, endpoint-scoped)
Settlement: 402Proof

Python SDK

from squeezeos_sdk import SqueezeOSClient
import os

client = SqueezeOSClient(xrpl_seed=os.environ["AGENT_XRPL_SEED"])
verdict = client.council("IWM")           # auto-pays 0.10 RLUSD, caches token
print(verdict["verdict"]["directive"])    # "BUY (IGNITION)"

Endpoint Pricing

Architecture

Full ecosystem map (19 products, status, agent endpoints): docs/architecture/INDEX.md

Agent Request
    │
    ▼
[MCP / REST]  ─── /mcp (JSON-RPC 2.0) or /api/* (REST)
    │
    ▼
[402Proof]    ─── HMAC-SHA256 token verify (pure CPU, no network)
    │
    ▼
[OracleEngine]─── aggregates 8 engines into one directive
    ├─ GammaFlowEngine    — gamma flip + dealer positioning
    ├─ SMLEngine          — fractal cascade depth 0–3
    ├─ BattleEngine       — multi-timeframe consensus
    ├─ OptionsIntelligence— sweep + whale detection
    ├─ VPINEngine         — order flow toxicity
    ├─ DarkPoolAxis       — dark print directional bias
    ├─ MeanReversionEngine— Ornstein-Uhlenbeck regime
    └─ IWM_ODTE_Engine    — 0DTE gamma/parity scoring
    │
    ▼
[Data Layer]  ─── Tradier (options) → Alpaca → Polygon → Alpha Vantage
    │
    ▼
[XRPL]        ─── Payments · URIToken notarization · Ghost Layer routing

Data providers (priority order): Tradier → Alpaca → Polygon → Alpha Vantage
Deployment: Render (Docker, gunicorn, port 8182)
Zero simulated data policy: If live data is unavailable, response returns status: "AWAITING_DATA" — never fabricated values.

Ecosystem

Agent Credit Bureau

FICO-style 300–850 score built from cryptographic XRPL spend history. Zero custody. Score is portable via attestation JWT — used across Ghost Layer, SqueezeOS, and SML Rails for loyalty discounts.

• Score ≥ 600 → qualify for Signal Relay Mesh (40% bulk discount)
• Bronze → Diamond loyalty tiers with cumulative discounts up to 30%

GET https://four02proof.onrender.com/v1/bureau/score/{wallet}

Discovery Files

Local Development

cp .env.example .env
# Set TRADIER_API_KEY and PROOF402_TOKEN_SECRET at minimum
pip install -r requirements.txt
python core/app.py
# or: gunicorn "core.app:create_app()"

Health check: GET /api/status

License

MIT — see LICENSE