How to Get BUFF163 Price History with an API

What CSPriceAPI provides

BUFF163 (NetEase) is one of the main Chinese marketplaces. Prices track YouPin closely; divergence between BUFF and YouPin history often signals arb windows.

Quick start

curl -s "https://api.cspriceapi.com/v1/prices/buff163/history?market_hash_name=AK-47%20%7C%20Redline%20(Field-Tested)" \
  -H "x-api-key: YOUR_API_KEY"

Compare BUFF vs YouPin on the same item:

curl -s "https://api.cspriceapi.com/v1/prices/youpin/history?market_hash_name=AK-47%20%7C%20Redline%20(Field-Tested)" \
  -H "x-api-key: YOUR_API_KEY"

Hourly (intraday, max ~100 days):

curl -s "https://api.cspriceapi.com/v1/prices/buff163/history?market_hash_name=AK-47%20%7C%20Redline%20(Field-Tested)&resolution=hourly" \
  -H "x-api-key: YOUR_API_KEY"

Current snapshot alongside history pull:

curl -s -X POST "https://api.cspriceapi.com/v1/prices/latest" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"items":["AK-47 | Redline (Field-Tested)"],"markets":["buff163","youpin"]}'

Reading history correctly

Hourly vs daily. Use resolution=hourly for short-range intraday charts (max ~100 days per request). Use resolution=daily (default) for backtests and trend analysis over 1+ year. Listing prices are sampled roughly every hour and rolled into daily closes.

Listing history ≠ sales. History tracks listing-price snapshots over time, not every completed trade. Confirm big moves with market sales.

Doppler: Request history per market_hash_name. Phase-specific items may need phase-aware names or separate handling — see Doppler guide.

Missing days: Gaps mean no snapshot stored — interpolate carefully in ML models.

Cross-market CN analysis

Plot BUFF and YouPin on the same chart (convert to one currency):

Align dates.
spread = buff_price - youpin_price in CNY.
When spread exceeds historical 90th percentile → arbitrage bot candidate.

No-code charts

Open Browse Skins
Search item → open skin detail
View multi-market history chart (BUFF, YouPin, Western markets)

Faster than building matplotlib for ad-hoc research.

Backtest sketch

# Load buff + youpin history JSON arrays
# For each day: signal = buff_close - youpin_close
# Simulate: buy cheap market, sell expensive (minus fees)
# Requires: fee model, FX if reporting USD

Pair with historical CS2 guide for other markets.