# AI use cases for Tick Stock Panel (shy3130/tick-stock-panel) Repo: [shy3130/tick-stock-panel](https://github.com/shy3130/tick-stock-panel) | Use cases: 3 | Page: https://dedupedai.com/use-cases/shy3130-tick-stock-panel/ TSP (tick-stock-panel) is a self-hosted quant workbench for China A-shares that runs in one Docker container with local Parquet and DuckDB storage. It covers custom and AI generated strategies, factor research and mining, backtests with T+1, fees and slippage, intraday alerts, an AI chat assistant, 65 Open API endpoints with token scopes and an MCP server with 12 tools. The README says it is not investment software and does no AI stock picking. Not financial advice: paper trade first, and remember you can lose money. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Self-host an A-share quant workbench in one container: https://dedupedai.com/use-cases/shy3130-tick-stock-panel/self-host-an-a-share-quant-workbench/ - Backtest A-share strategies with T+1, fees and slippage: https://dedupedai.com/use-cases/shy3130-tick-stock-panel/backtest-a-share-strategies-with-t1-and-fees/ - Connect an AI client to your quant workbench over MCP: https://dedupedai.com/use-cases/shy3130-tick-stock-panel/connect-an-ai-client-to-tsp-over-mcp/ ## How to self-host an A-share quant workbench in one container with Tick Stock Panel Repo: [shy3130/tick-stock-panel](https://github.com/shy3130/tick-stock-panel) | Difficulty: Beginner | Time: 20 minutes | Basis: From the README Pull the published image and run strategies, scans and dashboards locally with your data on your disk. **Who it is for:** A developer studying China A-shares who does not want cloud platforms holding their data. **The problem:** Patched together scripts and paid terminals make it hard to keep data, strategies and backtests consistent. **The outcome:** A local web workbench with dashboard, screener, backtest, mining and monitoring pages, data stored as local Parquet. ### Steps 1. Install Docker 2. Run the one-line docker run command from the README quick start, which pulls the ghcr.io/shy3130/tick-stock-panel image and mounts a local data folder 3. Open the workbench on localhost at the port given in the quick start 4. Copy .env.example to .env if you need a data source key or LLM settings ### Prompt to paste into your agent ```text Deploy shy3130/tick-stock-panel locally with the published Docker image from its README quick start, keep data in a mounted folder, and tell me which pages load and which data source mode it is in. ``` ### Caveat It targets China A-shares only and the README says it is for learning and research, not an investment or market watching tool. Not financial advice: paper trade first, and remember you can lose money. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/shy3130-tick-stock-panel/self-host-an-a-share-quant-workbench/ ## How to backtest A-share strategies with T+1, fees and slippage with Tick Stock Panel Repo: [shy3130/tick-stock-panel](https://github.com/shy3130/tick-stock-panel) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Run factor, strategy and minute level backtests that follow A-share T+1 rules and real costs. **Who it is for:** A quant learner whose simple backtests ignore A-share settlement rules and trading costs. **The problem:** Backtests that skip T+1, fees and slippage look far better than any real result. **The outcome:** Backtests on the same data path as live strategy runs, with factor attribution and no future data leaking in. ### Steps 1. Start the workbench as above 2. Build or import a strategy in the strategy engine, or let the AI generate one 3. Run a backtest with T+1, fees and slippage switched on 4. Review factor attribution and the backtest views before changing parameters ### Prompt to paste into your agent ```text In shy3130/tick-stock-panel, create a strategy for on A-shares, backtest it over with T+1, fees and slippage, and report return, drawdown and the factors that explain it. ``` ### Caveat Factor mining results are searched out of sample and published explicitly, but past fits can still fail live. Not financial advice: paper trade first, and remember you can lose money. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/shy3130-tick-stock-panel/backtest-a-share-strategies-with-t1-and-fees/ ## How to connect an AI client to your quant workbench over MCP with Tick Stock Panel Repo: [shy3130/tick-stock-panel](https://github.com/shy3130/tick-stock-panel) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README Expose the workbench's data and tools to an AI client through its MCP server and scoped API tokens. **Who it is for:** A developer who wants Claude, Codex or another MCP client to query their local quant data. **The problem:** Building bots on top of a dashboard usually means scraping pages. **The outcome:** An AI client calling 12 MCP tools on your local workbench, limited by the token scope you choose. ### Steps 1. Run the workbench and create an API token with the narrowest scope you need 2. Set up the bundled mcp-server, which runs over stdio 3. Add it to your MCP client config 4. Ask the client a data question and check the tool calls it made ### Prompt to paste into your agent ```text Connect shy3130/tick-stock-panel's mcp-server to with a read only token, list the available tools, and answer . Show me which tools were called. ``` ### Caveat By default the token gateway only accepts local or LAN traffic, so keep it that way unless you know why to open it. Not financial advice: paper trade first, and remember you can lose money. Risk, credentials: API tokens grant access to your workbench, so give each client the lowest scope and keep the gateway local only. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/shy3130-tick-stock-panel/connect-an-ai-client-to-tsp-over-mcp/