How Barfinex Works
Signals context API — strategy context and triggers
How Barfinex provides HTF context (candles, bias, zones) and order-flow data for strategies and the Advisor. GET /signals/context/:symbol and response shape.
What this is for
Barfinex supports a context + trigger approach for strategies and the Advisor:
- Higher-timeframe (HTF) candles (e.g. D1, H4, H1) provide context: regime, bias, levels, zones of interest.
- Intraday order flow and order book provide entry timing: absorption, order book imbalance, sweep and reclaim, pullback with confirmation.
- Risk and exit can be derived from HTF zones (e.g. stop beyond a zone, target at mid or opposite boundary).
The platform exposes this via the signals context API and related event streams so detectors and the Advisor can work with a consistent view of context and triggers.
Provider API: GET /api/signals/context/:symbol
This endpoint returns a normalized context for a symbol: candles (D1/H4/H1), order-flow aggregates, order-book aggregates, and optional context features (bias, regime, levels, zones).
Query parameters
| Parameter | Description | Default (typical) |
|---|---|---|
connectorType | Connector (e.g. binance) | binance |
marketType | Market (spot, futures, margin) | futures |
daysD1 | Days of D1 history | 365 |
daysH4 | Days of H4 history | 180 |
daysH1 | Days of H1 history | 90 |
candlesMode | FULL — return candle arrays; DELTA — omit candles if key matches | FULL |
candlesKey | Cache key from previous response; used with candlesMode=DELTA to avoid resending unchanged candles | — |
Example request (full context)
curl "http://localhost:8080/api/signals/context/BTCUSDT?connectorType=binance&marketType=futures&daysD1=365&daysH4=180&daysH1=90"
Example request (delta — avoid resending candles)
curl "http://localhost:8080/api/signals/context/BTCUSDT?connectorType=binance&marketType=futures&candlesMode=DELTA&candlesKey=<lastCandlesKey>"
Response shape (main parts)
candles
- cacheKey — Key you can send back as
candlesKeyfor delta mode. - h1, h4, d1 — Arrays of candles (time, open, high, low, close, volume). Omitted or replaced with only
cacheKeywhencandlesMode=DELTAand key matches.
orderFlow (aggregates)
Summary metrics for triggers (not tick-by-tick): e.g. delta ratio, CVD, absorption score, trade count, average/max trade size, buy/sell volume, delta, VWAP.
orderBook (aggregates)
- Best bid/ask, mid, spread, spread %, imbalance, bid/ask wall (e.g. in USD).
contextFeatures (optional)
Compact HTF context computed by the provider: bias (e.g. LONG/SHORT/NEUTRAL), regime (e.g. range/trend), levels (htfHigh, htfLow, htfMid), and zones (e.g. HTF_LOW, HTF_HIGH with range and timeframe). Used by the Advisor and detectors as a quick summary; candles remain the full source when needed.
How detectors and Advisor use it
- Detectors — Can use context (bias, zones) and trigger logic (order flow, order book) in a pipeline: e.g. no trade against HTF bias, entry only inside a zone, risk/exit from HTF range.
- Advisor — Receives context features and aggregates (and optionally candle summaries) in the decision payload; full candle arrays are not required for the LLM but are available in the detector/context pipeline when needed.
For full API details, see Provider API reference. For architecture, see Barfinex architecture.