Roadmap

The Backtesting API is the first module of a broader infrastructure stack. The long-term goal is to build complete, programmable infrastructure for quantitative intelligence and market execution — covering research, validation, deployment, and monitoring.

Every module follows the same API-first philosophy: accessible via HTTP, composable, and designed for programmatic integration rather than manual operation.

• Server-side backtest execution via HTTP API.
• JSON-based strategy DSL with composable inputs, conditions, scoring, and decision rules.
• Long and short strategies via configuration.direction.
• Signal-based exits (decision.exit), alongside stop-loss/take-profit.
• All timeframes from 5M to 1D.
• 50 curated pairs on Coinbase.
• Full metrics: trades, win rate, expectancy R, PnL, profit factor, condition diagnostics.
• Pay-as-you-go pricing based on candles processed.
• Console for API key management, usage, and billing.

A hosted Model Context Protocol server so AI agents can call the Backtesting API directly, without a custom integration. Same underlying API, same credits and usage — just a different way in for agent-based workflows.

• Hosted at mcp.backtest.emidlabs.com (Streamable HTTP) — add it to any MCP-compatible client config.
• Tools: submit_backtest, get_backtest_result, list_available_assets.
• strategy-dsl-spec resource so an agent knows the exact JSON shape to generate.
• Independent rate limiting: your account’s real quota passes through unchanged; a separate, loose IP-based guard protects the MCP process itself.

Full reference: MCP Server docs.

Subscribe a validated strategy to live market data and get a signal the moment its entry or exit condition fires on a real, closed candle — the same DSL and evaluation engine as the Backtesting API, running continuously instead of over a historical range. See Live Execution API docs for the full reference.

Composable signal distribution beyond a single subscription's own log — routing signals across multiple destinations and pipeline stages.

• Signal distribution: streams or multi-endpoint fan-out, beyond per-subscription polling.
• Signal pipelines: compose signals across subscriptions with filters, combiners, and routing rules.

• New markets: equities, futures, and additional crypto exchanges.
• New data types: tick data, order book snapshots, funding rates.
• Portfolio management: multi-strategy backtesting with capital allocation simulation.
• Multi-strategy orchestration: coordinate signals and execution across strategies.
• Automated quantitative pipelines: scheduled research, parameter optimization, and reporting.