A native C++20 terminal packing 37 AI agents and 100+ data feeds

What it does

Fincept Terminal v4 is a native C++20 desktop application that wraps quantitative finance tools, real-time market data, and algorithmic trading into a single binary. It uses Qt6 for the interface and embeds Python to run analytics like DCF models, portfolio optimization, and risk metrics. The app connects to over 100 data sources and 16 brokers, covering equities, crypto, fixed income, and derivatives.

The interesting bit

The project squeezes an entire trading stack—QuantLib modules, LLM agents, WebSocket feeds, even maritime tracking—into a desktop client that runs locally rather than in a browser. The 37 AI personas named after famous investors sit alongside genuine multi-provider LLM support and a local inference option, suggesting the authors care about keeping data on-premise.

Key highlights

• Native C++20/Qt6 UI with embedded Python analytics engine
• 100+ data connectors (Polygon, Kraken, FRED, IMF, World Bank, AkShare, etc.)
• Real-time trading via WebSocket with 16 broker integrations including IBKR, Alpaca, and Indian platforms like Zerodha
• 37 AI research agents across trading, economic, and geopolitical frameworks with local LLM support
• Node editor for visual automation pipelines and MCP tool integration

Caveats

• Build dependencies are pinned to exact versions (CMake 3.27.7, Qt 6.8.3, Python 3.11.9, specific compiler versions); the README warns that deviating may produce unstable binaries
• Docker support is explicitly limited to Linux/X11 and labeled CI-only; Windows and macOS Docker are not supported
• The “Global Intelligence” features (maritime tracking, satellite data) are listed but the README does not clarify data sources or coverage depth

Verdict

Worth a look if you want a local, hackable trading workstation with Python analytics baked in. Skip it if you need a lightweight web dashboard or a plug-and-play server that builds against arbitrary dependency versions.