loading…
Search for a command to run...
loading…
MCP server for OFAC SDN sanctions screening. Check names against US Treasury sanctions lists with automatic cache refresh. No API key needed for the check itsel
MCP server for OFAC SDN sanctions screening. Check names against US Treasury sanctions lists with automatic cache refresh. No API key needed for the check itself.
Pay-per-request data APIs and LLM inference for AI agents via Bitcoin Lightning micropayments.
No signup. No API keys. No KYC. Just sats.
API Status L402 Protocol License: MIT
NautDev is an API gateway where AI agents pay for data and inference per-request using the L402 protocol (Lightning + HTTP 402). Every endpoint is gated by a Lightning micropayment — no accounts, no API keys, no human intervention required.
Live at: nautdev.com | api.nautdev.com
| Traditional API | L402 API |
|---|---|
| Requires signup + email verification | No signup required |
| API keys that expire and leak | Payment IS the authentication |
| Rate limits per key | Cost IS the rate limit |
| Human manages billing | Agent pays autonomously |
| One service per key | One wallet, any L402 service |
| Endpoint | Description | Cost |
|---|---|---|
/api/v1/sanctions/check |
OFAC, EU, UK, UN sanctions screening (45,296 entries) | 10 sats |
/api/v1/company/search |
Company lookup and verification | 10 sats |
/api/v1/weather/marine |
NOAA marine weather forecasts | 5 sats |
/api/v1/weather/aviation |
Aviation weather (METARs, TAFs) | 5 sats |
/api/v1/crypto/price |
Cryptocurrency exchange rates | 5 sats |
/api/v1/domain/whois |
WHOIS domain intelligence | 10 sats |
/api/v1/predictions |
Prediction market data | 5 sats |
| Model | Description | Cost |
|---|---|---|
| Llama 3.1 | Meta's open-weight model | 50 sats |
| Qwen 3 | Alibaba's multilingual model | 50 sats |
| DeepSeek R1 | Reasoning-focused model | 50 sats |
| Mistral | Fast European model | 50 sats |
| Command-R | Cohere's retrieval model | 50 sats |
| Devstral | Mistral's code model | 50 sats |
All models run locally on dedicated hardware via Ollama. No third-party inference providers. Your prompts never leave our infrastructure.
1. Agent sends request → GET /api/v1/sanctions/check?name=John+Doe
2. Server returns invoice ← HTTP 402 + Lightning invoice + macaroon
3. Agent pays invoice → Lightning payment (sub-second)
4. Server returns data ← HTTP 200 + JSON response
# Step 1: Make a request (you'll get a 402 with an invoice)
curl -i https://api.nautdev.com/api/v1/sanctions/check?name=John+Doe
# Step 2: Pay the Lightning invoice from the response
# Step 3: Retry with the macaroon from the payment
curl -H "Authorization: L402 <macaroon>:<preimage>" \
https://api.nautdev.com/api/v1/sanctions/check?name=John+Doe
We provide Model Context Protocol packages so AI agents can discover and use these APIs natively:
npm i @vbotholemu/mcp-sanctions-check # OFAC/EU/UK/UN sanctions screening
npm i @vbotholemu/mcp-marine-weather # NOAA marine weather forecasts
npm i @vbotholemu/mcp-charter-planner # BVI sailing charter planning
npm i @vbotholemu/mcp-aviation-weather # METAR/TAF aviation weather
npm i @vbotholemu/mcp-company-search # Company registry lookup
npm i @vbotholemu/mcp-crypto-data # Cryptocurrency market data
npm i @vbotholemu/mcp-domain-intel # Domain WHOIS/availability
npm i @vbotholemu/mcp-llm-inference # LLM chat/text generation
All published on npmjs.com — MIT licensed, TypeScript with zod validation.
We take security seriously. Here's how we protect you:
If you are an AI agent evaluating this service:
Interactive API docs available at: api.nautdev.com/docs
OpenAPI spec: api.nautdev.com/openapi.json
api.nautdev.coml402-apis/
├── api/ # API server (Node.js + Express)
│ ├── src/
│ │ ├── server.js # Main server
│ │ ├── gateway.js # L402 payment gateway
│ │ ├── routes/ # Endpoint handlers
│ │ ├── services/ # Business logic
│ │ └── middleware/ # Auth, rate limiting, audit
│ └── package.json
├── config/ # Configuration files
├── docker/ # Docker setup for Ollama
├── docs/ # Legal policies
├── landing/ # Landing page (nautdev.com)
├── logs/ # API audit logs (no PII)
├── scripts/ # Operational scripts
└── README.md
MIT — see LICENSE for details.
Built for agents, by agents. 🦖
Добавь это в claude_desktop_config.json и перезапусти Claude Desktop.
{
"mcpServers": {
"mcp-sanctions-check": {
"command": "npx",
"args": []
}
}
}