Ecosystem¶
fincore is one member of the cloudQuant quantitative-research ecosystem. The projects below are designed to work together: fincore supplies the metrics and performance analytics, the backtrader family covers strategy execution, authoring tooling, and web workflows.
| Project | Focus | Description |
|---|---|---|
| backtrader | Strategy execution | Professional Python algorithmic-trading framework for backtesting and live trading, actively maintained by cloudQuant. |
| backtrader-skills | Authoring tooling | Offline author/review/test product for the backtrader fork: turns local datasets and typed StrategySpec v1 into pytest strategies or three-file bundles, reviews candidates without importing them, and runs approved candidates in isolated child processes. |
| backtrader-mcp | LLM integration | Local-first MCP server for building and running reproducible backtrader strategies: immutable datasets, private strategy drafts, and bounded subprocess runs with durable status and reports. Offline and backtest-only. |
| backtrader_web | Web platform | "AI for Investor": a web-based full-cycle backtrader strategy management platform (Vue 3 + FastAPI) covering research, strategy generation, backtesting analysis, paper trading, live execution, and data management. |
| backtrader-agent | Agent runtime | Offline-first strategy-authoring agent runtime: content-addressed data storage, canonical strategy specifications, static review, hash-bound approvals, and a fixed child-process execution profile with recoverable session provenance. |
| fincore | Analytics core | This repository: unified Python toolkit for financial metrics (150+), performance analysis, backtesting support, AI-driven insights, and multi-database/data source integration. |
How the pieces fit together¶
fincore ────────────── metrics, risk, attribution, reports
▲
│ numeric evidence
│
backtrader ─────────── strategy execution (backtest & live)
▲ ▲
│ │
backtrader-skills ──┴─ author/review/test tooling for strategies
backtrader-mcp ─────── MCP server exposing reproducible strategy workflows to LLM tools
backtrader-agent ───── offline strategy-authoring agent runtime
backtrader_web ─────── web platform connecting research, backtest, paper, and live trading
The backtrader family exchanges data through typed contracts (strategy specifications, immutable datasets, bounded subprocess runs), so each project stays independently installable and offline-first where applicable.