tick-stock-panel is a self-hosted Docker workbench that unifies A-share stock screening, monitoring, backtesting and review through one consistent data pipeline with pluggable data sources. It stores data locally in Parquet and includes an optional AI conversational assistant over local data.
Project overview
Its capability routing matrix lets any dataset switch between pluggable data sources without changing indicator or backtest semantics, addressing the common pain of locked-in data sources and inconsistent metrics across screening, backtesting and monitoring — all in a single-container, zero-ops deployment.
Project type
Workflow
Use cases
Data Analysis
Deployment
Refer to project documentation
License
MIT
Best for
A-share quantitative researchers and learners who want screening, monitoring, backtesting and review unified in one self-hosted workbench without local Python or Node setup
Users who need to swap data sources without rebuilding indicators or backtests
Key capabilities
A capability routing matrix lets dataset types route independently across pluggable data sources, so sources can be swapped without changing metric or backtest semantics.
Limitations and risks
This is a personal open-source project; the author prioritizes sponsor support.
AI features are disabled unless an AI key is configured, and the assistant fails closed with providers that lack tool calling.
Backtest results do not represent future returns; the project is intended for learning and quantitative research only.
Codex CLI mode lets the container read host Codex login credentials; it should only be enabled in trusted local environments.
Getting started
Run a single docker command using the prebuilt GHCR image — no local Python or Node required for the recommended path: docker run -d --name tsp -p 3018:3018 -v ${PWD}/data:/app/data ghcr.io/shy3130/tick-stock-panel:latest, then open http://localhost:3018.