Infrastructure

APIs, Data & System Architecture

AlgoBoy is built as an algorithmic infrastructure platform, not a collection of dashboards. Our systems support real algorithmic work: research, testing, deployment, and monitoring.

Core Premise: Everything we expose is intended to be deterministic, reproducible, and suitable for integration into automated pipelines. We do not provide "tips"; we provide the fuel for autonomous systems.

API Philosophy & Design

Our API architecture is governed by strict principles designed to prevent the common failures found in retail-grade data providers. We prioritize machine-readability over human convenience.

1. Determinism First

Given the same inputs, parameters, and version, an endpoint will always return the same output. No hidden state. No adaptive behavior. This is critical for valid backtesting.

2. Explicit Over Convenient

We prefer verbose, explicit parameters over "smart" endpoints that guess intent. If a value is derived, it is labeled. If a transformation is applied, it is inspectable.

3. Infrastructure, Not Advice

Endpoints return data, signals, classifications, and metrics. We do not issue trades. All outputs are strictly inputs for your decision engine.

Data Schemas & Structure

AlgoBoy uses strict, versioned JSON schemas. We employ strong typing and predictable field names across all market categories (Equities, Betting Exchanges, Futures). Breaking schema changes always result in a new API version, never a silent update.

GET /v1/markets/{id}/snapshot

{
  "market_id": "1.2059382",
  "timestamp": "2026-01-21T14:30:00Z",
  "status": "OPEN",
  "data_quality": "L1_VERIFIED",
  "runners": [
    {
      "selection_id": 559201,
      "symbol": "TRAP_4",
      "state": {
         "last_price": 4.2,
         "volume_traded": 15402.50,
         "implied_prob": 0.238
      }
    }
  ]
}
        

Signal Endpoints

Unlike raw data, our Signal Endpoints provide derived metrics. These are not "black boxes"; they are calculated indicators based on published methodologies. Common signals include Regime Classifications, Liquidity Stress flags, and Relative Strength metrics.

Design Rule: If a signal cannot be independently validated or backtested, it does not belong in production.

POST /v1/signals/regime-detection

# Python Client Example
import algoboy

client = algoboy.Client(api_key="your_key")
signal = client.get_regime(
    market="US_FUTURES_ES",
    window="4h",
    model="volatility_breakout_v2"
)

print(signal.classification)
# Output: "HIGH_VOLATILITY_EXPANSION"
        

Historical vs. Live Contexts

We make a rigid distinction between historical data (for research) and live data (for execution). Mixing these implicitly is a primary cause of lookahead bias in algorithmic modeling.

Historical Endpoints

Designed for Backtesting & Training
Guaranteed Forward-Only logic
No Lookahead Leakage
Stable Reconstruction

Live / Near-Live Endpoints

Designed for Execution Pipelines
Explicit Latency Expectations
Defined Failure Behavior
Real-time Websocket Feeds

Versioning & Reproducibility

Reproducibility is non-negotiable in science and finance. Every endpoint is versioned. If we publish a research paper referencing an API, we cite the specific API version used. This ensures that you can reconstruct the exact conditions of the study 12 months later.

Older versions remain available where feasible to support long-running strategy evaluations.

Rate Limits & Usage Models

AlgoBoy APIs are designed for serious usage, not scraping. We utilize predictable, tiered rate limits to protect system integrity.

Research Access: High bandwidth for historical data, limited live polling.
Production Access: Low latency, high frequency live polling, websocket priority.
Institutional: Custom agreements for high-throughput data ingestion.

Ready to Build?

Join quantitative researchers and data engineers building on deterministic infrastructure. Get your API keys and start testing today.

Request API Access