Market intelligence, one JSON call per ticker.
A composite verdict that fuses trend stage, Point & Figure technicals, insider activity and the macro regime into a 0–100 score with a plain-English rationale — plus every underlying signal as its own endpoint, webhooks on transitions, and an MCP server so AI assistants can call it directly.
What you get
Composite verdict
GET /v1/verdict/{ticker} — one 0–100 score, a stable accumulate · hold · distribute · avoid enum, per-factor sub-scores, confidence, and the reasoning spelled out. One call replaces six.
Stage analysis
Which of the four trend stages a ticker is in — basing, advancing, topping or declining — from a 30-week moving-average model, plus a walk-forward historical weekly series up to ten years back — backtest-safe, no lookahead.
Point & Figure and breadth
P&F buy/sell signals and relative strength versus the benchmark, and sector bullish-percent breadth across a curated universe.
Macro regime
Risk-on / neutral / risk-off derived from credit spreads, rates, the dollar and positioning (FRED and CFTC data), with the point-by-point detail behind the call.
SEC filings
Per-transaction Form 4 insider activity — who, role, buy or sell, shares, price — and recent 8-Ks, straight from EDGAR, summarised per ticker.
Webhooks and the macro calendar
Subscribe to stage transitions, P&F signal changes, regime changes and insider clusters; signed deliveries with retries. Upcoming FOMC and CPI dates on /v1/calendar.
Built for agents and LLM tools
Self-describing responses
Every response carries as_of, a freshness enum and suggested_recheck_seconds, so an agent knows how much to trust a value and when to ask again. Signals ship with a quotable rationale.
MCP server
An open-source Model Context Protocol server wraps the read endpoints as tools for Claude Desktop, Claude Code and other MCP clients. Each call is metered against your own key — nothing hidden.
Honest billing
Only successful responses cost credits. A symbol the feed doesn't know, or an upstream outage, is never billed. Duplicate tickers in one request count once.
Pricing
One credit per signal per ticker. The composite verdict and the historical stage series cost 5 credits per call. Paid plans keep working past the included credits and bill the excess per credit at month end; the Free plan stops at its allowance.
Free / Dev
- 1,000 credits a month
- 2 requests per second
- Hard cap — upgrade to keep going
- No webhooks
- All signals, MCP server, API reference
Pro
- 50,000 credits a month
- 10 requests per second
- Overage $1.00 per 1,000 credits
- Webhooks included
- All signals, MCP server, API reference
Business
- 500,000 credits a month
- 50 requests per second
- Overage $0.80 per 1,000 credits
- Webhooks included
- All signals, MCP server, API reference
Prices in USD, billed by Stripe. Change or cancel any time from the customer portal (POST /v1/billing/portal). Paid plans start on the Free tier and upgrade in one click from the signup page or via POST /v1/billing/checkout.
Quick start
# 1. Get a key at https://api.vanoe.ai/signup, then:
curl -H "Authorization: Bearer sk_live_…" \
"https://api.vanoe.ai/v1/verdict/AAPL"
# → {"data": {"ticker": "AAPL", "score": 70.6, "verdict": "accumulate",
# "confidence": 0.74, "rationale": "Stage 2 advance …", …},
# "as_of": "2026-09-08", "freshness": "end_of_day",
# "suggested_recheck_seconds": 86400, "request_id": "…"}
# Several tickers in one call (1 credit each):
curl -H "Authorization: Bearer sk_live_…" \
"https://api.vanoe.ai/v1/signals/stage?tickers=AAPL,MSFT,NVDA"
# What you've used this period:
curl -H "Authorization: Bearer sk_live_…" "https://api.vanoe.ai/v1/usage"
Full reference with every endpoint, parameter and error code at /docs. Errors are structured with stable codes and a request_id; rate limits return 429 with Retry-After.
Where the data comes from
Macro series from the Federal Reserve (FRED), positioning from the CFTC Commitments of Traders report, insider and 8-K filings from SEC EDGAR, and end-of-day market prices from our price feed. Signals are computed from that data on a daily and weekly cadence — this is not a real-time quote service. Everything is informational and can be wrong, late or unavailable; see the terms.