What is TSP (tick-stock-panel)?
TSP is an open-source workbench for learning and quantitative research on Chinese A-shares. It connects strategy creation, factor analysis, backtesting, monitoring and post-market review through a shared data pipeline. Data-source plugins route datasets according to provider capabilities, while custom fields can be exposed alongside built-in analysis pages.
The application can run in a Docker container with local Parquet storage and DuckDB queries, without an external database service. Desktop packages are also documented for Windows and Apple Silicon Macs. Its extension architecture supports custom strategies, signals, data sources and interface pages.
Optional AI features support strategy drafts, stock analysis and conversational queries against the deployment's data. The assistant exposes tool-call details and requests confirmation for actions by default. External scripts and AI clients can access selected capabilities through a scoped API and an MCP server; the project explicitly excludes AI stock recommendations and presents itself as a research tool rather than investment software.
Key capabilities
- Strategy research Creates custom strategies and signals, with optional AI-generated drafts and screening across Chinese A-shares.
- Backtesting and factors Supports strategy, factor and minute-level studies with settlement constraints, fees, slippage and out-of-sample evaluation.
- Market monitoring Combines auction, intraday and deviation alerts with configurable rules, voice announcements and Feishu notifications.
- Data extensibility Routes datasets between plugin-based providers and adds custom fields and analysis pages.
- AI assistance Queries local datasets through inspectable tool calls and uses confirmation cards for actions by default.
- Programmatic access Provides scoped API tokens, server-sent events and an MCP bridge for external tools and AI clients.
Where TSP (tick-stock-panel) fits
TSP fits a local research workflow in which a developer or technically capable researcher imports A-share data, defines screening conditions, tests strategies and monitors resulting signals. Custom data tables and page extensions allow additional datasets and research views to sit alongside the built-in workbench.
Before deployment, users should verify provider credentials, dataset coverage and provider terms, configure access controls, and preserve and back up the persistent data volume during upgrades. AI assistance requires a configured provider with tool-calling support. Remote API access needs explicit network configuration, and the documented macOS desktop package is limited to Apple Silicon.
Install
docker compose up --buildReviews of TSP (tick-stock-panel)
Be the first to share your experience with TSP (tick-stock-panel).
Write a review
Loading review form…
Alternatives to TSP (tick-stock-panel)
HummingbotHummingbot is an Apache 2.0 framework for automated crypto trading strategies. It runs locally or in Docker, connects to centralized and decentralized exchanges and offers scripts, controllers, executors, paper trading and the hbot CLI.
KronosKronos is a family of open-source decoder-only models pre-trained on candlestick data from over 45 global exchanges, with Python inference and fine-tuning.