Alpaca Trading API - Technical Reference

Overview

Commission-free trading platform with REST API and WebSocket streaming. Developer-friendly alternative to Java-based platforms like Interactive Brokers. Supports stocks, options, and cryptocurrency trading through HTTP requests.

Configuration

Account Setup

• Minimum Balance: $0 to open account
• Pattern Day Trading: Requires $25,000 minimum equity for unlimited day trading
• Margin Trading: $2,000 minimum equity required
• Verification: 24-48 hours for account approval
• International: Accepted from most countries, US securities only

API Rate Limits

• Free Tier: 200 API calls per minute
• Paid Tier: 10,000 API calls per minute ($99/month)
• Critical Warning: Free tier limit hit in 30 seconds when rebalancing 50-stock portfolio

Authentication

• Separate API key pairs for paper and live trading
• OAuth 2.0 support for third-party applications
• IP restriction capability for API keys
• TLS 1.3 encryption with 2FA support

Market Data Configuration

Data Tiers

Tier	Cost	Coverage	Latency
IEX (Free)	$0	3% trading volume, all symbols	15-minute delay
SIP Data	$99/month	All exchanges	Sub-20ms real-time

WebSocket Streaming

• Connection Reliability: Drops during high volatility when most needed
• Reconnection: Python SDK reconnection logic broken in newer versions
• SSL Issues: ssl.SSLEOFError common in Python 3.11+
• Batch Updates: Automatic for high-volume symbol tracking

Critical Failure Modes

API Connection Issues

• WebSocket Disconnections: Every 20 minutes during volatility
• SSL Errors: ConnectionResetError: [Errno 104] indicates server issues
• Rate Limiting: 403 Forbidden: rate limit exceeded even at 50 requests/minute
• WSL2 Specific: WebSocket drops every 20 minutes due to Microsoft networking

Trading Reality vs Paper Trading

• Performance Gap: 15% backtested → 8% paper → 3% live returns typical
• Slippage Impact: Paper trading fills at midpoint, live gives market scraps
• Spread Costs: 0.05% spread on small-caps destroys returns on $50k positions
• Partial Fills: Not simulated in paper trading

Production Deployment Issues

• Memory Leaks: Common in long-running bots
• Dependency Hell: pandas >=1.2.0 numpy >=1.20.0 version conflicts
• JSON Validation: KeyError: 'timestamp' during market open - always validate responses

Resource Requirements

Time Investment

• Setup: 10 minutes (ideal) to 2 hours (typical)
• First Trade: ~1 hour for basic implementation
• Production Ready: Weeks to months for robust system

Expertise Costs

• Basic Trading: HTTP request knowledge sufficient
• Production Systems: Requires error handling, reconnection logic, risk management
• Debugging: Environmental issues (Python versions, SSL, WebSocket) require system administration skills

Financial Costs

• Trading: $0 commissions + regulatory fees (FINRA TAF $0.000166/share max $8.30, SEC $0.0000278/dollar)
• Data: $99/month for real-time consolidated data
• Infrastructure: Self-hosted or cloud deployment costs

Platform Comparison

Feature	Alpaca	Interactive Brokers	Charles Schwab	TD Ameritrade
Setup Complexity	Low	High (Java hell)	Medium	Medium
API Reliability	Good (volatility issues)	Poor (random disconnects)	Slow responses	Medium
Rate Limits	200/min (10k paid)	50/sec	120/min	120/min
International Markets	No	Yes	No	No
Crypto Trading	Built-in	No	No	No
Best Use Case	US retail algo trading	Complex international	Conservative trading	Legacy TD users

Success Patterns

Effective Strategies

• Timeframe: Minute-to-hour execution (avoid sub-minute scalping)
• Portfolio Size: <50 stocks to avoid rate limits on free tier
• Order Types: Standard market/limit orders most reliable
• Risk Management: Implement custom position limits beyond platform defaults

Working Configurations

• Python SDK: Stable with Python 3.8-3.10, issues with 3.11+
• WebSocket Handling: Build custom reconnection with exponential backoff
• Error Recovery: docker system prune -a && docker-compose up for nuclear reset
• Dependency Management: Pin specific versions to avoid SSL/numpy conflicts

Critical Warnings

What Documentation Doesn't Tell You

• Scalping Limitation: 1-second strategies get rate-limited immediately
• WebSocket Reliability: Expect disconnections during high volatility
• Paper Trading Deception: Performance gap is significant and predictable
• Spread Impact: Small-cap stocks have spreads that devour algorithmic returns
• SSL Certificate Issues: Newer Python versions break randomly

Breaking Points

• Rate Limits: Hit at 50 requests/minute despite 200/minute limit
• Memory Usage: Long-running bots develop memory leaks
• Network Stability: WSL2 users experience predictable 20-minute disconnections
• Market Volatility: All connection issues worsen during high-volume periods

Risk Controls Reality

• Platform Safeguards: Basic position limits and PDT compliance only
• Custom Implementation Required: Day trading limits, sector concentration, stop-losses
• Margin Calls: Trigger at worst possible timing
• Account Protection: FINRA/SIPC coverage for platform failures, not trading losses

Decision Criteria

Choose Alpaca When:

• Building US-focused retail trading bots
• Need crypto integration
• Want simple HTTP API over Java complexity
• Commission-free structure is priority
• Development speed matters over advanced features

Avoid Alpaca When:

• Requiring international markets
• High-frequency trading (sub-minute)
• Need ultra-reliable connections during volatility
• Complex multi-asset strategies
• Institutional-grade risk management required

Migration Considerations

• From IBKR: Significant simplification but feature reduction
• From Schwab/TD: Similar features, better developer experience
• To Production: Expect 6-month development cycle for robust system
• Cost Analysis: Free tier sufficient for learning, $99/month for serious trading

Troubleshooting Guide

Common SSL/Connection Fixes

# Nuclear option for connection issues
docker system prune -a && docker-compose up

# Dependency reset
rm -rf node_modules && npm install
# Or for Python
pip uninstall alpaca-trade-api && pip install alpaca-trade-api==1.4.1

Error Code Meanings

• ssl.SSLEOFError: Python version compatibility issue
• ConnectionResetError [Errno 104]: Alpaca server instability
• 403 Forbidden rate limit: Actual usage exceeds documented limits
• KeyError: 'timestamp': JSON response validation failure
• HTTP 404 WebSocket: Complete environment reset required

Support Resources Quality

• GitHub Issues: Responsive, actual solutions
• Community Forum: Slow but helpful for trading-specific problems
• Discord: 24/7 community support, best for urgent issues
• Documentation: Accurate examples that work when copied
• Status Page: Essential bookmark for production deployments