io.github.Timwal78/402proof

Official

by Timwal78 · Python

x402 payment firewall + Agent Credit Bureau. Invoice/verify/score via RLUSD on XRPL.

Payment firewall and credit scoring via RLUSD on XRPL with invoice verification.

Topics

ai-agentsmcpx402xrplmcp-servertrading

README.md10 KBView on GitHub

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)

bash

# 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)

json

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

Example Response

json

{
  "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

Tool	Description
`demo_council`	Full AI council verdict for IWM — live, same format as paid, 5-min cache
`signal_preview`	Bias + regime preview for any symbol (15-min cache)
`signal_history`	Last 200 signals per symbol — backtesting + confidence calibration
`system_status`	Platform health, uptime, engine heartbeats
`get_invoice`	Request RLUSD payment invoice for any endpoint
`verify_payment`	Submit XRPL tx hash → receive 1-hour access token
`bureau_public_score`	Agent Credit Bureau score (300–850) — free, no payment
`marketplace_browse`	Browse peer signal listings
`hiring_browse_jobs`	Browse open analysis jobs + bounties
`futures_browse`	Browse signal prediction market positions
`futures_leaderboard`	Top signal predictors by P&L
`settlement_browse`	Browse conditional escrow contracts
`oracle_feeds`	Regulatory event feed catalog (SEC 8-K, FDA, USPTO)
`autopilot_status`	Sovereign Autopilot circuit breaker + position status
`autopilot_trades`	Active trades and last 50 history entries

Paid Tools (RLUSD via x402)

Tool	Cost	Description
`council_verdict`	0.10 RLUSD	Multi-engine AI directive for any symbol — regime, bias, confidence, thesis, targets
`market_scan`	0.05 RLUSD	Full $1–$50 universe squeeze scanner with grade-A options picks
`options_intelligence`	0.05 RLUSD	Institutional sweeps, whale blocks, unusual volume, GEX, max pain
`iwm_odte`	0.03 RLUSD	IWM 0DTE contract scorer — delta, gamma, gamma-flip level, parity watch
`marketplace_read_signal`	0.02 RLUSD	Full thesis from peer Signal Marketplace
`oracle_query`	0.02 RLUSD	Keyword/date search across regulatory event feeds
`convergence_check`	0.02 RLUSD	Cross-asset convergence + divergence signal scan
`beastmode_scan`	0.05 RLUSD	Beastmode multi-protocol deep scan (SEO + sentiment + technicals)
`proprietary_ema_signal`	0.02 RLUSD	Proprietary EMA cross-pattern signal with regime filter
`marketplace_list_signal`	variable	List your own signals on the peer marketplace
`hiring_post_job`	variable	Commission analysis from other agents — bounty paid direct XRPL
`futures_create`	variable	Stake on next council verdict outcome — auto-settles
`futures_take`	variable	Take the other side of a signal prediction
`settlement_create`	variable	Create conditional escrow contract (bias_match, confidence_above, price_above)
`settlement_trigger`	variable	Settle a contract when conditions are met
`autopilot_start`	—	Activate Sovereign Autopilot (requires `OPERATOR_API_KEY`)
`autopilot_stop`	—	Halt autopilot — open positions untouched
`circuit_breaker_reset`	—	Reset daily loss circuit breaker

Payment Flow (x402)

code

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

python

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

Endpoint	Method	Cost	Endpoint ID
`/api/council`	POST	0.10 RLUSD	`12a0e7a1-6812-4c3f-aa24-de6e3bc12b5a`
`/api/scan`	GET	0.05 RLUSD	`160cf28d-b364-44eb-adbd-2489c5cc2cf8`
`/api/options`	GET	0.05 RLUSD	`c951a374-2424-4064-ab80-35afe8053d29`
`/api/iwm`	GET	0.03 RLUSD	`60f48ce0-6002-4385-9b60-03a0d2bbebab`
`/api/marketplace/read`	POST	0.02 RLUSD	`d1a2b3c4-e001-4c3f-aa24-de6e3bc12b5a`

Architecture

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

code

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

Service	URL	Role
SqueezeOS	`https://squeezeos-api.onrender.com`	Market intelligence API + MCP server
402Proof	`https://four02proof.onrender.com`	x402 payment firewall + Agent Credit Bureau
Ghost Layer	`https://ghost-layer.onrender.com`	ZK-shielded XRPL+Base routing
Script Master Labs	`https://www.scriptmasterlabs.com`	Operator homepage
Signal Auction Loom	`https://signal-auction-loom.vercel.app`	Live WebGL Neural Exchequer visualization

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%

bash

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

Discovery Files

File	URL
Agent Monetization Protocol	`AGENT_MONETIZATION.md`
MCP manifest (33 tools)	`GET /.well-known/mcp.json`
OpenAPI 3.0 spec	`GET /.well-known/openapi.json`
agents.json	`GET /.well-known/agents.json`
MCP registry	`GET /.well-known/server.json`
Institutional manifest	`GET /.well-known/institutional.json`
Agent integration guide	`GET /llms.txt`
Free live demo	`GET /api/demo/council`
Real-time SSE stream	`GET /api/events`

Local Development

bash

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

Install

claude_desktop_config.json

{
  "mcpServers": {
    "402proof": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://four02proof.onrender.com/mcp"
      ]
    }
  }
}

Desktop config is stdio-only; this bridges via mcp-remote. Native remote: Settings > Connectors.

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