Data Sources - ClearView Portal

ClearView never generates data from memory. Every number you see — in dashboards, chat responses, and automated reports — is fetched from a primary source. This page is the complete reference for where the data comes from.

Crypto Data

Binance (Spot)
Binance (Futures)
DefiLlama

What: BTC and ETH spot prices, 24-hour price changes, and historical candlestick data (klines) going back up to 9 years.Used by: Macro Agent, Quantitative Agent, Structure Agent, Technical Agent

Endpoint	Data	Update Frequency
`/api/v3/ticker/24hr`	Current price, 24h change, volume	~1 minute
`/api/v3/klines`	Historical OHLCV candlesticks	On request (up to 1,000 candles per call)

Binance Spot is the source for all crypto price data in the platform. When you see a BTC or ETH price anywhere — in a dashboard, a chart, or an AI response — it comes from here.

What: Derivatives data from Binance USDT-margined futures — the most liquid crypto derivatives venue.Used by: Derivatives Agent, Signals Agent

Endpoint	Data	Update Frequency
`/fapi/v1/fundingRate`	Funding rate history (every 8h settlement)	Every 8 hours
`/fapi/v1/openInterest`	Current open interest	~1 minute
`/futures/data/globalLongShortAccountRatio`	Long/short account ratio	Every 5 minutes
`/futures/data/takerlongshortRatio`	Taker buy/sell volume ratio	Every 5 minutes

Funding rates settle every 8 hours (00:00, 08:00, 16:00 UTC). Between settlements, the rate shown is an estimate of the next settlement value.

What: Stablecoin market data — total supply, supply by chain, and individual stablecoin metrics.Used by: Macro Agent

Endpoint	Data	Update Frequency
`/stablecoins`	All stablecoins with supply and market cap	~30 minutes
`/stablecoins/chains`	Stablecoin supply broken down by blockchain	~30 minutes

Stablecoin supply is a widely watched indicator of capital entering or leaving the crypto ecosystem.

Traditional Finance Data

EODHD
FRED (Federal Reserve)
Finnhub

What: A comprehensive financial data provider covering equities, indices, forex, economic calendars, earnings, fundamentals, news, insider trades, and stock screening.Used by: TradFi Agent, Technical Agent, Advisor Agent

Category	Examples	Update Frequency
Real-time quotes	Stock/ETF prices, index levels	~1 minute
Fundamentals	P/E, market cap, revenue, earnings	~4 hours
Economic calendar	FOMC, CPI releases, GDP	~30 minutes
Earnings calendar	Earnings dates and estimates	~30 minutes
News & sentiment	Company and market news	Real-time
Insider transactions	SEC filings	Daily
Screener	Filter stocks by financial criteria	On demand

EODHD is the primary source for all traditional finance data in ClearView.

What: Official economic data from the Federal Reserve Bank of St. Louis — the authoritative source for US macroeconomic indicators.Used by: TradFi Agent, Macro Agent

Series	Indicator	Update Frequency
DGS2, DGS5, DGS10, DGS30	Treasury yields (2Y, 5Y, 10Y, 30Y)	Daily
WALCL	Fed balance sheet (total assets)	Weekly
M2SL	M2 money supply	Monthly
CPIAUCSL	Consumer Price Index (CPI)	Monthly
PCEPI	Personal Consumption Expenditures (PCE)	Monthly
UNRATE	Unemployment rate	Monthly
GDP	Gross Domestic Product	Quarterly
DTWEXBGS	Trade-weighted US Dollar Index (DXY proxy)	Daily

FRED data updates on the schedule of the underlying economic releases. GDP is quarterly, CPI is monthly, treasury yields are daily.

What: Market status — whether exchanges are currently open or closed.Used by: TradFi Agent

Data	Coverage	Update Frequency
Exchange status	US markets (NYSE, NASDAQ)	~1 minute

Currently limited to US market status on the free tier.

AI Models

The AI layer processes data fetched by agents — it does not generate data.

Model	Role	What it does
Claude Sonnet 4.5	Query planner	Reads your question and decides which agents to invoke. Fast and precise.
Claude Opus 4.6	Response synthesizer	Takes all agent data and writes the final analysis with context and cross-references.

Both models are from Anthropic. The planner runs on every query (deciding the strategy), and the synthesizer runs when assembling the final response or generating Deep Synthesis reports.

What’s NOT included

On-chain data (blockchain metrics)

Currently being rebuilt using Google BigQuery public datasets. When available, this will provide 23 proprietary metrics including MVRV, NUPL, HODL waves, realized cap, NVT ratio, and more — all computed from raw blockchain data rather than third-party providers like Glassnode or CryptoQuant.

ETF flow data

No free, reliable API source exists for real-time ETF flow data (including Bitcoin and Ethereum spot ETFs). This may be added when a suitable data provider becomes available.

Cross-exchange derivatives

All derivatives data comes from Binance only. Binance handles approximately 60% or more of global crypto futures volume, making it a reliable proxy for the broader derivatives market. Aggregating across smaller, less liquid exchanges would add noise without meaningful signal.

Crypto options data

Crypto options (Deribit, etc.) are not currently covered. Equity options are partially available via EODHD for major US stocks, but coverage varies by ticker.

Real-time liquidation data

Binance retired the public REST endpoint for liquidation data. Real-time liquidations will be available in a future update using WebSocket streaming, which provides the data as it happens rather than after the fact.

Caching and Freshness

All data passes through a Redis cache layer. Cache durations are set to match how frequently each data type actually changes:

Data Type	Cache Duration	Rationale
Real-time prices	1 minute	Prices move continuously
Funding rates	A few minutes	Settles every 8 hours, estimates update frequently
Calendar and earnings	30 minutes	Event schedules change infrequently
Fundamentals	4 hours	Quarterly data, changes rarely
FRED macro indicators	6 hours	Monthly or quarterly releases

When a data source is temporarily unavailable (API down, rate limited), the platform shows a clear error state with a retry option. You will never see stale data presented as current. If the cache expires and the source is unreachable, the affected section shows an error message instead of outdated numbers.

Every AI response includes a Sources section at the bottom listing exactly which endpoints provided each data point. If you ever want to verify a number, the source attribution tells you where to look.

Reference

​Crypto Data

​Traditional Finance Data

​AI Models

​What’s NOT included

​Caching and Freshness

Crypto Data

Traditional Finance Data

AI Models

What’s NOT included

Caching and Freshness