gprophet 0.1.4

Official Python SDK for the G-Prophet AI Stock Prediction API

G-Prophet Python SDK

Official Python client for the G-Prophet AI Stock Prediction API.

Installation

pip install gprophet

To also use the MCP Server (for AI agents like Claude, Kiro, etc.):

pip install gprophet[mcp]

Quick Start

from gprophet import GProphet

client = GProphet(api_key="gp_sk_your_key_here")

# Check balance
print(client.balance())

# Get a stock quote
quote = client.quote("AAPL", market="US")
print(f"AAPL: ${quote['price']}")

# AI prediction
pred = client.predict("AAPL", market="US", days=7)
print(f"Predicted: ${pred['predicted_price']} ({pred['direction']})")

# CN/A-share prediction uses G-Prophet2026V2 in auto mode, with V1 as fallback.
cn_pred = client.predict("600519", market="CN", days=7, algorithm="gprophet2026v2")
print(f"600519: ¥{cn_pred['predicted_price']} ({cn_pred['direction']})")

# Technical analysis
tech = client.technical("AAPL", market="US")
print(f"Signal: {tech['overall_signal']}")

# Batch quotes
quotes = client.batch_quote(["AAPL", "GOOGL", "MSFT"], market="US")
for q in quotes["results"]:
    if q["success"]:
        print(f"{q['symbol']}: ${q['price']}")

# AI stock analysis (async with auto-polling)
analysis = client.analyze_stock("AAPL", market="US", wait=True)
print(analysis["analysis"]["summary"])

MCP Server (for AI Agents)

After installing with pip install gprophet[mcp], a gprophet-mcp command is available.

See MCP Server documentation for setup details.

API Reference

System

• client.health() — Health check
• client.info() — API metadata, pricing, supported markets

Account

• client.balance() — Points balance and quota info
• client.usage(days=7) — Usage statistics

Predictions

• client.predict(symbol, market, days, algorithm) — AI price prediction
• client.compare(symbol, market, days, algorithms) — Multi-algorithm comparison

Supported prediction algorithms include auto, gprophet2026v1, gprophet2026v2 (CN/A-shares only), lstm, transformer, random_forest, and ensemble. For client.compare(...), the default CN algorithm list includes gprophet2026v2; other markets do not expose V2.

Market Data

• client.quote(symbol, market) — Real-time quote
• client.batch_quote(symbols, market) — Batch quotes
• client.history(symbol, market, period) — Historical OHLCV
• client.search(keyword, market, limit) — Search symbols

Technical Analysis

• client.technical(symbol, market, indicators) — Technical indicators & signals

Sentiment

• client.fear_greed(days) — Crypto Fear & Greed Index
• client.market_overview(market) — Market overview

Analysis

• client.analyze_stock(symbol, market, locale, wait) — AI stock analysis (58 pts)
• client.analyze_comprehensive(symbol, market, locale, wait) — Multi-agent analysis (150 pts)
• client.task_status(task_id) — Check async task status

Error Handling

from gprophet import GProphet
from gprophet.client import GProphetError, RateLimitError, InsufficientPointsError

client = GProphet(api_key="gp_sk_...")

try:
    result = client.predict("AAPL", market="US")
except RateLimitError as e:
    print(f"Rate limited, retry after {e.retry_after}s")
except InsufficientPointsError as e:
    print(f"Need {e.required} points, have {e.available}")
except GProphetError as e:
    print(f"API error [{e.code}]: {e.message}")

Configuration

client = GProphet(
    api_key="gp_sk_...",
    base_url="https://www.gprophet.com/api/external/v1",  # Custom base URL
    timeout=60,       # Request timeout (seconds)
    max_retries=3,    # Auto-retry on rate limit
)

License

MIT