Market grounding for AI agents and quantitative research. One call returns the consolidated, factual context an agent needs before making a decision — technicals, fundamentals, insider flows, news sentiment, macro regime — instead of ten scattered sources.
curl -H "X-API-Key: YOUR_KEY" https://<host>/agent/context?symbols=NVDA,AMAT
{
"regime": { "label": "calm", "crash_score": 2, "vix": 17.4 },
"macro": { "...": "timestamped series" },
"global_markets": [ "... Asia / Europe / US context" ],
"tickers": {
"NVDA": { "technicals": {}, "fundamentals": {}, "news": {} }
},
"missing_data": { "AMAT": ["insiders"] }
}
| Access | How | Limit |
|---|---|---|
| API key | header X-API-Key (or ?api_key=) | daily quota per key |
| Anonymous | public demo paths only (/symbols, /chat) | rate-limited per IP |
Over-quota returns 429 {"error":"quota_exceeded"}; missing key on a data endpoint returns 401.
| Endpoint | What you get |
|---|---|
GET /agent/contextflagship |
The consolidated grounding payload for AI agents: regime, macro, sectors, and per-ticker
technicals + fundamentals + insiders + news sentiment + earnings, in one call.
?symbols=NVDA,AMD. Alias: /brain/context for the internal BrainPortfolio pipeline. |
GET /ticker/{symbol} | Everything on one symbol (see quickstart). |
GET /snapshot | Market snapshot: indices, VIX, breadth, regime summary. |
GET /sectors | 11 SPDR sectors, 5d/20d rotation, inflow/outflow ranking. |
GET /macro · GET /macro/{series} | Macro series (rates, curve, spreads, inflation…). |
GET /inflationnew | CPI, Core CPI, PCE and Core PCE year-over-year, computed from FRED levels, with distance to the Fed 2% target. |
GET /macro/central-banksnew | Fed, ECB, BoJ and BoE rate series when available. Missing regions are explicit. |
GET /fxnew | Major FX pairs and DXY with latest price and day change. |
GET /markets/regionalnew | Regional ETF rotation across US, Europe, Japan, China, Korea and EM. |
GET /crash | Multi-signal market stress score with per-axis detail. |
GET /movers | Notable movers of the day with context. |
GET /news · GET /sentiment | Market news with scored sentiment; aggregate sentiment view. |
GET /insiders | SEC Form 4 insider flows, 30-day windows, noise-filtered. |
GET /earnings | Upcoming earnings calendar with estimates. |
GET /read?url=new | Reads any article, research page or press release: main text extracted, cached, ready to ground an agent. The chat uses it automatically when you paste a URL. |
GET /search?q=new |
Full-text search over every document Helix has read (Fed, SEC, BLS, market press) — RAG-ready for agents. |
GET /symbols?q= | Symbol search / autocomplete. |
GET /status | Public uptime, collector freshness, row counts and data health. |
GET /freshness | Raw last successful run per collector. |
GET /health | Minimal liveness check. |
Interactive OpenAPI reference: /docs
LLM agents hallucinate numbers when they lack data. Helix exists to make that impossible: every field is collected from primary sources, timestamped, and returned as plain facts — no narrative, no advice. You reason; Helix grounds.
One call gives your agent the facts; you keep the reasoning. Provider-agnostic — the pattern is the same for Claude, GPT or a local model.
import requests, anthropic
sym = "NVDA"
ctx = requests.get(
f"https://<host>/agent/context?symbols={sym}",
headers={"X-API-Key": "YOUR_KEY"}).json()
msg = anthropic.Anthropic().messages.create(
model="claude-haiku-4-5",
max_tokens=300,
system=("You are a market analyst. Use ONLY the facts in "
"HELIX_DATA below. Never invent a number.
"
f"HELIX_DATA:
{ctx}"),
messages=[{"role": "user",
"content": f"What does the data say about {sym} right now?"}])
print(msg.content[0].text)
No RAG plumbing, no ten data vendors to stitch — the agent gets regime, macro, technicals, fundamentals, insiders and news in a single grounded blob.
250 requests/day, no card, shown once. You also get the plain-language Daily Brief when email delivery opens.
Paste your key, pick a ticker, see the live payload.
Data is informational only. Not financial advice. No warranty on availability during early access.