dpx-mcp
AI intelligence oracle + cross-border settlement rail. 10-layer Stability Oracle (climate, macro, FX, ESG, supply chain) with x402 pay-per-call data API. USDC/EURC settlement on Base at 1.385% all-in. GENIUS Act + MiCA + Basel III compliant.
README
DPX MCP Server
AI-native infrastructure for institutional finance. DPX exposes a 10+ layer intelligence oracle and a compliance-grade cross-border settlement rail as a single MCP server - any AI agent can read macro signals, score ESG counterparty risk, and execute settlements with no human in the loop.
13 tools | Base mainnet | GENIUS Act + MiCA + Basel III | x402 pay-per-call
What DPX is
DPX is two things built as one system:
1. An AI intelligence layer - a self-adjusting oracle that synthesizes 32+ real-time signals across climate, macroeconomics, FX, energy transition, supply chain, earth systems, and ESG. The oracle is not a static data feed: it recalibrates tier weights weekly against prediction outcomes (Platt scaling), stores scenario fingerprints in a Vectorize index, and runs a Claude AI synthesis layer on top of all nine data tiers. The intelligence endpoints are a standalone data product: pay-per-call in USDC via x402, no API key required.
2. A settlement rail - cross-border USDC/EURC settlement on Base mainnet at 1.385% all-in, versus 2-5% for SWIFT or 5.4% for Stripe. Every settlement is ESG-scored, oracle-gated, and compliant with GENIUS Act (US), MiCA (EU), and Basel III by default. The same AI oracle that generates the intelligence briefings also gates whether a settlement executes.
Intelligence endpoints (standalone data product)
Deployed at https://intelligence.untitledfinancial.com. Pay per call in USDC on Base via x402 - no API key, no subscription.
| Endpoint | Price | Cache | What it returns |
|---|---|---|---|
GET /intelligence |
$0.10 | live | AI narrative synthesis across all 32+ signals - score, tier, alerts, forward outlook |
GET /intelligence/climate |
$0.25 | 24h | Precipitation anomalies across 10 agricultural zones vs 1972-74 baseline - commodity exposure, food inflation signals, basket currency implications |
GET /intelligence/earth-systems |
$0.50 | 48h | Planetary health dashboard (Earth Health Index 0-100) - CO‚ÇÇ/CH‚ÇÑ/temp/sea ice vs pre-industrial, proximity to 9 climate tipping points |
GET /intelligence/macro-stress |
$0.15 | 1h | Credit regime classification - IG/HY OAS spreads, TED, VIX, C&I lending tightening - 2-6 week lead signals for FX and commodity moves |
GET /intelligence/supply-chain |
$0.25 | 6h | Lane bottleneck scoring - NY Fed GSCPI + live Rhine/Mississippi/Panama/Great Lakes water levels - goods inflation lead signal |
GET /intelligence/energy-transition |
$0.25 | 24h | Renewable share, grid carbon intensity, fossil demand curve - structural energy price context |
GET /intelligence/esg/:address |
$0.25 | 6h | Entity-level ESG score from GLEIF + SEC EDGAR + EPA ECHO + OSHA - counterparty due diligence |
Data sources: Open-Meteo ERA5, NASA GISTEMP, NOAA GML, NSIDC, FRED (BAMLC0A0CM, BAMLH0A0HYM2, TEDRATE, VIXCLS, DRTSCILM), NY Fed GSCPI, WSV Pegelonline, USGS NWIS, EIA, GLEIF, SEC EDGAR, EPA ECHO, OSHA.
x402 example - no API key:
import { withPaymentInterceptor } from 'x402-fetch';
const fetchWithPayment = withPaymentInterceptor(fetch, wallet);
const data = await (await fetchWithPayment('https://intelligence.untitledfinancial.com/intelligence/macro-stress')).json();
// ‚Üí { stressIndex: 42, regime: "LATE_CYCLE", leadSignals: { fxImplication: "USD_STRENGTH_RISK", ... } }
MCP tools (for AI agents)
Install once, use from any MCP-compatible host (Claude Desktop, Cursor, n8n, custom agents).
| Tool | Description |
|---|---|
get_manifest |
DPX protocol capabilities and contract addresses |
get_quote |
Binding fee quote (300s TTL) - all-in rate with live ESG adjustment |
get_esg_score |
Live ESG score for a wallet address - hourly from 6 institutional sources |
get_reliability |
Oracle stability status - STABLE / CAUTION / UNSTABLE |
get_oracle_status |
Full 10+ layer Stability Oracle v9.0 output - all tier scores, alerts, chaos regime flag, AI briefing |
get_fee_schedule |
Complete fee table with volume tiers and competitive benchmarks |
verify_fees |
Confirm off-chain quote matches on-chain DPXSettlementRouter |
compare_to_competitors |
DPX vs Stripe, Wise, SWIFT, bank wire - dollar savings per transaction |
get_rail_status |
Live health of local payment rails (PIX, SEPA, FedACH, CHAPS, UPI, PromptPay) |
settle |
Execute a settlement - cross-border or domestic, sandbox or live |
get_settlement_status |
Look up a settlement by ID - full audit record |
get_investment_context |
Structured investment memo for AI due diligence agents |
get_intelligence |
MPP-gated macro intelligence briefing from the Stability Oracle |
Quick Start
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"dpx": {
"command": "npx",
"args": ["-y", "@untitledfinancial/dpx-mcp"]
}
}
}
Restart Claude Desktop. All 13 tools are immediately available.
Remote HTTP (no install)
{
"mcpServers": {
"dpx": {
"type": "http",
"url": "https://mcp.untitledfinancial.com/mcp"
}
}
}
Smithery
npx @smithery/cli install @untitledfinancial/dpx-mcp --client claude
Example flows
AI agent settlement workflow:
Agent: settle $2M USD ‚Üí EUR for 0xRecipient...
get_oracle_status ‚Üí 10+ layer signal check - score 81, STABLE, no chaos regime
get_esg_score ‚Üí counterparty ESG tier confirmed (score 74 ‚Üí 0.15% ESG fee)
get_quote ‚Üí $27,700 all-in at 1.385%, quoteId valid 300s
verify_fees ‚Üí on-chain fee matches off-chain quote ‚úì
settle ‚Üí executed on Base mainnet, txHash + IPFS receipt returned
Intelligence query (x402, no MCP):
AI reads macro-stress ‚Üí regime LATE_CYCLE, USD_STRENGTH_RISK flag ‚Üí agent reduces FX exposure
AI reads climate ‚Üí wheat CRITICAL in South Asia ‚Üí commodity desk alerted
AI reads earth-systems ‚Üí AMOC proximity HIGH ‚Üí long-horizon risk model updated
Oracle architecture
The 10+ layer Stability Oracle v9.0 underpins both the intelligence product and the settlement rail:
| Layer | Signal |
|---|---|
| 0 - Climate & Environmental | 30-90 day lead indicators |
| 1 - Commodities & Energy | Spot + structural signals |
| 2 - Macroeconomic | 4 independent sources per indicator |
| 3 - Currency & FX | 4 cross-validated APIs |
| 4 - Basket Verification | On-chain Chainlink + 3 FX APIs |
| 5 - Bond Yields | FRED yield curve |
| 6 - Infrastructure + Conflict | Dual sub-modules |
| 7 - Cross-Body Integration v8.0 | Geopolitical, capital flows, tech/AI supply chain |
| 8 - USD Structural Health v9.0 | 25+ signals, 10% composite blend |
| 9 - AI Synthesis | Claude reasoning layer - degrades gracefully |
Adaptive layer: Tier weights recalibrate weekly against prediction outcomes via Platt scaling. Scenario fingerprints stored in Cloudflare Vectorize. Autonomous policy execution (basket rebalance, fee adjust) up to $10M notional with 23h cooling period and circuit breaker.
Compliance
GENIUS Act (US) | MiCA (EU) | EU SFDR/CSRD (transaction-level ESG) | Basel III | FCA/PSR (UK) | FATF Travel Rule
Live contracts (Base mainnet)
| Contract | Address |
|---|---|
| DPXSettlementRouter | 0x7d2b0Cea5A2d19369548F59C6B8EEe9Fe3495c97 |
| DPXToken | 0x7A62dEcF6936675480F0991A2EF4a0d6f1023891 |
| ESGCompliance | 0x7717e89bC45cBD5199b44595f6E874ac62d79786 |
Verified on Sourcify. Full list: get_manifest
Links
- Homepage: mcp.untitledfinancial.com
- Intelligence API docs: INTELLIGENCE_API.md
- npm: @untitledfinancial/dpx-mcp
- Smithery: smithery.ai/servers/untitledfinancial/dpx-mcp
- Docs: docs.untitledfinancial.com
License
MIT © Untitled Financial
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。