Repolode
Submit
shy3130

shy3130/tick-stock-panel

Self-hosted A-share quant workbench for building strategies, backtesting, live monitoring and AI-assisted analysis, with open API and MCP access.

Open-source alternative to:Tonghuashun iFinDTongdaxin

View on GitHubWebsite
Stars
6,190
Forks
1,473
Open issues
33
Last commit
3 hours ago

What is it?

TSP is a self-hosted quantitative research workbench for China's A-share market. It runs as a single Docker container with Parquet and DuckDB file storage, so no external database is required. Users can write or AI-generate screening strategies, run factor and strategy backtests with T+1, fees and slippage, configure intraday and auction anomaly monitors with push notifications, and review market sentiment and sector rankings.

Data sources are pluggable. The project routes six dataset types by each source's declared capabilities, so switching providers does not change indicator or backtest definitions. Supported sources listed in the project include TickFlow, fuyao (Tonghuashun REST), and TdxQuant, and custom YAML sources can be added.

For developers, the core capabilities are exposed through a token-gated Open API with six permission scopes and an MCP server with 12 tools, so external scripts and AI clients can read data and trigger backtests. An optional AI assistant answers questions over local data and requires an OpenAI-compatible endpoint with tool calling.

Key features

  • 01Custom strategy screening across the A-share market with AI-assisted strategy generation
  • 02Factor and strategy backtesting with T+1, fees, slippage and minute-level replay
  • 03Intraday, auction and deviation anomaly monitoring with rule-based alerts and Feishu push
  • 04Pluggable data sources with per-capability routing
  • 05Open API with token scopes and an MCP server for external tools and AI clients
  • 06Single-container Docker deployment storing all data locally

Use cases

  • Individual investors who want to test A-share screening rules and backtest them locally
  • Quant researchers building custom factors and validating them out of sample
  • Developers who want to pull panel data into their own scripts or bots through the Open API
  • Users wiring an AI client to their local market data via MCP

Quick start

  1. 1

    Run the project as a single Docker container, following the README quick-start section

Things to know

  • Focused on the Chinese A-share market only.
  • The README says it is for local quantitative research and learning, not as investment or market-watching software.
  • Market data depends on pluggable third-party data sources that you must connect yourself.

Topics

#a-stock#quantitative-trading#backtesting#stock-screener#fastapi#polars#duckdb#self-hosted#ai-agent#aigc

FAQ

What is tick-stock-panel?
Self-hosted A-share quant workbench for building strategies, backtesting, live monitoring and AI-assisted analysis, with open API and MCP access. TSP is a self-hosted quantitative research workbench for China's A-share market. It runs as a single Docker container with Parquet and DuckDB file storage, so no external database is required. Users can write or AI-generate screening strategies, run factor and strategy backtests with T+1, fees and slippage, configure intraday and auction anomaly monitors with push notifications, and review market sentiment and sector rankings.
Is tick-stock-panel free to use? What license does it have?
tick-stock-panel is open source under the MIT license. Check the license terms before using it in commercial products.
Is tick-stock-panel still maintained?
Actively maintained. The last commit to tick-stock-panel was on Oct 21, 2026.
What can I use tick-stock-panel for?
Individual investors who want to test A-share screening rules and backtest them locally Quant researchers building custom factors and validating them out of sample Developers who want to pull panel data into their own scripts or bots through the Open API Users wiring an AI client to their local market data via MCP
How do I install tick-stock-panel?
Run the project as a single Docker container, following the README quick-start section
Where can I get tick-stock-panel?
The source code and installation instructions are on GitHub: https://github.com/shy3130/tick-stock-panel

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects