Quickstart
Get your first causal prediction in 3 minutes.
1. Install the SDK
terminalbash
pip install abel-cap2. Get your API key
Sign up at abel.ai/pricing to get a free API key (1,000 calls/month included).
3. Make your first prediction
quickstart.pypython
import abel
client = abel.Client(api_key="sk-your-key-here")
# Predict BTC price using causal Markov blanket
prediction = client.predict("BTCUSD_close", horizon=48)
print(prediction)
# → { value: 63420, ci_lower: 61800, ci_upper: 65100,
# mb_variables: ["ETHUSD_close", "DXY", "SP500", ...] }4. Explain causal drivers
explain.pypython
# What's causally driving BTC?
drivers = client.explain("BTCUSD_close", depth=2, cross_domain=True)
print(drivers)
# → { parents: [
# { name: "ETHUSD_close", influence: "STRONG_POSITIVE", lag: "SHORT_TERM" },
# { name: "DXY", influence: "MODERATE_NEGATIVE", lag: "SHORT_TERM",
# cross_domain: true }
# ],
# cross_domain_chains: [
# "Fed_Funds_Rate →[τ=5h]→ DXY →[τ=2h]→ BTCUSD_close"
# ] }5. Simulate a causal intervention
intervene.pypython
# What happens if the Fed raises rates 50bp?
effect = client.intervene(
treatment="Fed_Funds_Rate",
outcome="BTCUSD_close",
treatment_value=0.5
)
print(effect)
# → { effect: -4.2%, ci: [-2.1%, -6.8%],
# causal_chain: "Fed →[τ=5h]→ DXY →[τ=2h]→ BTC",
# method: "do-calculus via PCMCI graph" }Next steps
- → Full API Reference — all 8 CAP primitives
- → Schema API — let your LLM discover the causal world model
- → MCP Server setup — use Abel from Claude or GPT
- → Concepts — understand why Abel is fundamentally different