Degenscan Intel

One normalized feed of price-moving events from ~40 primary sources (SEC, Federal Reserve, Federal Register, ECB/BoE/BoJ, FTC/DOJ/FDA/CFTC/FCC, USGS, NOAA/NHC, Nasdaq halts, DefiLlama, Polymarket, press wires…), scored against an exposure graph into per-asset impacts. Deterministic, sub-200 ms, auditable (raw_ref links to the source document).

• MCP (streamable HTTP): POST https://degenscan-intel.onrender.com/mcp
• REST: https://degenscan-intel.onrender.com/v1/...
• Docs for machines: https://degenscan-intel.onrender.com/llms.txt
• Operator: Marbella Collins LLC · contact@degenscan.io · MIT

When to use which tool

Recommended loop (every 1–4 h, or before acting)

1. universe once per session → map your tickers to asset ids (NVDA, BTC, CL, US10Y, SPX…).
2. regime_snapshot → if nothing high-severity touches your book and venues are closed, stop here.
3. events_since(since="4h", universe=[...your book...], min_confidence=0.4) → candidate events.
4. For each asset with |bias| meaningful: impact_for(asset_id, since="24h").
5. Check tradable_now / next_open on the event before sending an order; if the venue is closed, queue for next_open.
6. If an impact looks wrong, explain(event_id) before acting; log the rationale.

Prediction markets

polymarket_context("Fed rate cut in October?") (or a slug/id) → current yes_prob, change_24h, and the primary-source events in the last 48h that match the question, with tier, corroboration, relevance. The signal is a fresh primary event the market has not repriced. Combine with explain(event_id) when the direction matters.

Reading the output

• direction: 1 supportive, -1 negative, 0 unclear. confidence (0..1) = source tier × severity/novelty × graph-path weight — a ranking signal, not a probability.
• severity (0..1) is event magnitude; novelty (0..1) drops for repeated/corroborating items.
• corroboration.count = how many independent sources reported it. Primary sources (SEC, Fed, USGS) originate events; media only corroborates.
• path shows the exposure-graph route (e.g. nat.quake → facility:TSMC-Fab18 → company:TSM → company:NVDA).
• universe_version is stamped on every response; use it to make backtests reproducible.

Paying

• 100 free calls/day per IP, then HTTP 402 with x402 v2 requirements (USDC on Base, eip155:8453). Any x402 client (@x402/fetch, x402-axios, Coinbase AgentKit) pays automatically.
• Prepaid key, no human: POST /v1/keys/x402/pack_1k (answer the 402 with USDC) → { api_key } with 1,000 calls, lifetime. Also pack_10k ($40) and pack_100k ($300). Check balance: GET /v1/keys/me with X-API-KEY.
• Card (for the human operator): GET /v1/plans → Stripe checkout → key.
• Send the key as header X-API-KEY on /v1/* or POST /mcp.

Examples

REST:

curl 'https://degenscan-intel.onrender.com/v1/events?since=4h&universe=NVDA,TSM,BTC&min_confidence=0.4'
curl 'https://degenscan-intel.onrender.com/v1/impact/MSTR?since=24h'
curl 'https://degenscan-intel.onrender.com/v1/graph/NVDA?depth=2'
curl  https://degenscan-intel.onrender.com/v1/regime

MCP tools/call:

{ "name": "events_since", "arguments": { "since": "4h", "universe": ["NVDA","BTC","CL"], "min_confidence": 0.4, "limit": 30 } }

Claude Desktop / Cursor config (remote MCP):

{ "mcpServers": { "degenscan-intel": { "url": "https://degenscan-intel.onrender.com/mcp", "headers": { "X-API-KEY": "<optional>" } } } }

Limits and honesty

• Coverage is US-equity-heavy (top-100 by volume + indices), 15 crypto, main commodities/FX/rates. Not every ticker exists — check universe.
• Impacts are heuristics over public events; there is no price data in the feed and no forecast. Information and analytics only — not investment advice.
• Sources marked best-effort in sources_status may go quiet; the feed degrades gracefully.