Elena & backtesting

Elena is QuantDesk’s dedicated backtest lifecycle agent. It is not MIKEY-AI (general protocol intelligence). Elena orchestrates dataset selection, strategy specs, simulation, walk-forward validation, and deploy gating.

Status

Component	Status	Notes
`elena-ai/` service (port 3006)	Shipped	Conductor + tool registry
`/api/v2/backtest/*`	Shipped	Job queue, manifests, deploy gate
Pro `BT` + Lite backtest flows	Shipped	Real APIs — no mock-only production path
Live LLM Conductor	Optional	Set LLM API key; `VITE_MOCK_ELENA=true` for UI-only dev
GPU optimization pool	Optional	CPU path default; `use_gpu: true` falls back with notice

Architecture

Browser → backend :3002 (/api/elena/chat, /api/v2/backtest/*)
              ↓
         elena-ai :3006 (Conductor tools)
              ↓
         Redis job queue → Python sim workers
              ↓
         databaseService (manifests, ledger, deploy bundles)

Elena never re-implements sim logic — it calls backend /api/v2/backtest/* as a client.
MIKEY has no tool that submits sim jobs directly.

Gateway routes

Route	Purpose
`POST /api/elena/chat`	Conductor chat (privacy-scrubbed proxy to Elena)
`GET /api/elena/health`	Elena service health via gateway

Backtest API (narrative)

Base: /api/v2/backtest. Full schemas in Swagger (/api/docs/).

Route	Purpose
`GET /datasets`	Browse oracle-history datasets (`dataset_id`, version, resolution)
`POST /runs`	Submit backtest job (requires fees, slippage, funding — zero-cost rejected)
`GET /runs/:id`	Manifest + metrics (Sharpe, Sortino, max DD, DSR, trade count)
`POST /runs/:run_id/deploy-gate/check`	Pass/fail vs thresholds (min trades, OOS Sharpe, max DD)
`POST /runs/:run_id/deploy-bundle`	Versioned artifact for paper → live `place_order_v2`
`POST /compliance/scan`	AST guardrail before queue (lookahead, blocklisted imports)
`POST /optimize`	Parameter sweeps (optional GPU queue)
`POST /walk-forward`	Rolling OOS validation

StrategyCardSchema

Developer output emits StrategyCardSchema JSON — SSOT for max_leverage, cost model refs, and Solana friction (priority fees + slippage floor). Deploy bundles extend this schema; there is no competing parallel format.

Deploy gate → flex publish

Complete backtest → run_id
deploy-gate/check passes (evaluates DSR, not raw Sharpe alone)
Optional paper/live deploy via scoped session keys
Publish flex card: POST /api/v2/social/flex-cards with run_id — see Verified flex charter

Local dev

# Backend + Redis required
cd backend && pnpm run start:dev

export QUANTDESK_BACKEND_URL=http://localhost:3002
cd elena-ai && pnpm dev   # port 3006

# Optional mock Conductor UI without LLM key
# frontend/.env: VITE_MOCK_ELENA=true

Fleet: bash scripts/services/start-all.sh from repo root (full fleet; docs site not auto-started).

V2 API endpoints — social, markets, data plane
Building on QuantDesk — place_order_v2 live deploy
Data plane overview — dataset dumps and candles
Verified flex charter

Overview

Getting started

Trading

Social

Data

Developers

Security and trust

FAQ

Changelog

Elena & backtesting API

Elena & backtesting

Status

Architecture

Gateway routes

Backtest API (narrative)

StrategyCardSchema

Deploy gate → flex publish

Local dev

​Elena & backtesting

​Status

​Architecture

​Gateway routes

​Backtest API (narrative)

​StrategyCardSchema

​Deploy gate → flex publish

​Local dev

​Related

Elena & backtesting

Status

Architecture

Gateway routes

Backtest API (narrative)

StrategyCardSchema

Deploy gate → flex publish

Local dev

Related