Twelve Data

Stocks, forex, and commodities price data. Traditional markets only — not for crypto.

Script Usage

This skill ships a single exports.py with all functions. Call it from a bash block:

python3 - <<'EOF'
import sys, json
sys.path.insert(0, "/data/workspace/skills/twelvedata")
from exports import twelvedata_price, twelvedata_quote, twelvedata_time_series

# Single quote
print(twelvedata_quote(symbol="AAPL"))

# Time series (last 30 daily candles)
series = twelvedata_time_series(symbol="AAPL", interval="1day", outputsize=30)
print(json.dumps(series.get("values", [])[:3], indent=2))
EOF

Available functions in exports.py: twelvedata_price, twelvedata_quote, twelvedata_time_series, twelvedata_eod, twelvedata_quote_batch, twelvedata_price_batch, twelvedata_search, twelvedata_stocks, twelvedata_forex_pairs, twelvedata_exchanges. Read exports.py directly when you need exact signatures.

Function Reference (signatures)

All functions are in exports.py. Symbols use TwelveData format (e.g. AAPL, EUR/USD, XAU/USD). Use prepost=True for pre/post-market data on US stocks.

Keyword → Tool Lookup

TwelveData vs CoinGecko — Boundary

MISTAKES — Read Before Calling

❌ MISTAKE 1: Using TwelveData for crypto

User: "BTC 价格"
❌ WRONG: twelvedata_quote(symbol="BTC/USD")
✅ RIGHT: coin_price(ids="bitcoin")  ← CoinGecko

❌ MISTAKE 2: Calling quote individually for multiple stocks

User: "AAPL MSFT GOOGL 现在什么价"
❌ WRONG: twelvedata_quote("AAPL"), twelvedata_quote("MSFT"), twelvedata_quote("GOOGL")  ← 3 calls
✅ RIGHT: twelvedata_quote_batch(symbols=["AAPL", "MSFT", "GOOGL"])  ← 1 call, up to 120 symbols

❌ MISTAKE 3: Forex without slash

❌ WRONG: twelvedata_quote(symbol="EURUSD")
✅ RIGHT: twelvedata_quote(symbol="EUR/USD")  ← always slash format

Also: USD/CNH not USDCNH, GBP/JPY not GBPJPY.

❌ MISTAKE 4: Expecting bid/ask from quote

❌ WRONG: "EUR/USD bid: 1.0850, ask: 1.0852"  ← quote only returns close
✅ RIGHT: "EUR/USD current price: 1.0851 (close/last price — bid/ask spread not available)"

❌ MISTAKE 5: Wrong interval format

❌ WRONG: twelvedata_time_series(interval="1D")  ← uppercase
✅ RIGHT: twelvedata_time_series(interval="1day")

Valid: 1min, 5min, 15min, 30min, 1h, 2h, 4h, 8h, 1day, 1week, 1month

❌ MISTAKE 6: Using full output when compact suffices

User: "AAPL 最近走势"
❌ WRONG: twelvedata_time_series(symbol="AAPL", interval="1day", outputsize="full")  ← 5000 candles
✅ RIGHT: twelvedata_time_series(symbol="AAPL", interval="1day", outputsize="compact")  ← 30 candles

Use full only when user explicitly needs deep history.

Symbol Reference

Commodities (verified)

Popular Forex Pairs

Use twelvedata_search to discover others.

Pre/Post-Market Data

twelvedata_quote(symbol="AAPL", prepost=true)
twelvedata_time_series(symbol="AAPL", interval="1min", prepost=true)

Returns premarket_change, premarket_change_percent, postmarket_change, postmarket_change_percent when available.

Output Sizes

• compact — Last 30 data points (default, faster). Use for "最近走势".
• full — Up to 5000 data points. Use for deep analysis / charting.

Proxy-Safe Usage

1. Agent tool calls: Always prefer twelvedata_* tools (this skill).
2. Platform code (skills/, tools/): use core.http_client.
3. Workspace scripts (bash): Do NOT call TwelveData directly. Use skill tools.

Compound Queries

Stock Comparison

1. twelvedata_quote_batch(symbols=["AAPL", "MSFT", "GOOGL", "TSLA"])
2. twelvedata_time_series(symbol="AAPL", interval="1day", outputsize="compact")  ← only for the one user cares most about

Macro Dashboard (stocks + forex + commodities)

1. twelvedata_quote_batch(symbols=["SPY", "QQQ"])          → US indices
2. twelvedata_quote_batch(symbols=["EUR/USD", "USD/JPY"])   → Forex
3. twelvedata_quote(symbol="XAU/USD")                       → Gold