Skip to content
Conduit

mcp server

portfolio-daily-tracker

Self-hosted portfolio journaling and agent-driven stock research: reproducible experiments, backtesting, durable jobs, API/CLI/MCP extensions, and multi-currency tracking.

Stepuuu/portfolio-daily-trackercommunitylistedunreviewednot running here yet
Use tool

Set up

Runs on your computer. Pick your agent and operating system.

This project does not publish an install method we can turn into a config. Follow the README instead.

README

synced Oct 14, 2026

📊 Portfolio Daily Tracker

Self-hosted portfolio journaling and agent-driven stock research

自托管投资组合记录与 Agent 股票研究工作台

Checks License: MIT Python 3.10+ React 18

📖 Documentation / 文档

🇬🇧 English · 🇨🇳 中文 · Feishu workbench / 飞书工作台

News / Versions

Version / 版本Date / 日期Updates / 更新摘要
V3.x2026-03-08† → 2026-09-14Portfolio tracking, AI chat and backtesting, expanded with a confirmed ledger, manual/Agent stock research and a Feishu card workbench.从组合追踪、AI 对话与回测，扩展到需确认的交易账本、手动与 Agent 股票研究及飞书卡片工作台。
V2 · Self-hosted engine / 自托管引擎2026-03-08†Moved tracking to Python with local JSON/CSV storage, multi-market quotes, risk metrics and automated reports.转为 Python 自托管，使用本地 JSON/CSV，支持多市场行情、风险指标与自动日报。
V1 · Google Sheets2025-09-25†Started with Google Sheets and Apps Script for daily snapshots, price updates, asset charts and monthly P&L.基于 Google Sheets 与 Apps Script，实现每日快照、行情更新、资产图表与月度盈亏记录。

† Historical milestones: V1 uses the earliest dated guide entry; V2 and V3 were first recorded together in the repository. These are not tagged release dates. † 历史里程碑：V1 日期取自最早的指南更新记录；V2 与 V3 首次收录于同一次仓库提交。这些日期不代表正式 Release 发布日。

Full changelog / 完整更新记录 · GitHub Releases · Roadmap / 后续计划

Highlights / 亮点

Feature功能
🌍Multi-market: A-shares, HK, US多市场：A 股、港股、美股
📈Sharpe, volatility, max drawdown夏普、波动率、最大回撤
🤖AI chat with GPT / Claude / DeepSeekAI 对话助手
🔬Manual and autonomous stock research手动实验与 Agent 自动股票研究
📊Transaction ledger and research journal交易账本与研究复盘
📉Strategy backtesting engine策略回测引擎
🔔Auto daily report → Feishu / Telegram每日自动推送日报

Try it with fictional data / 先体验演示数据

Requires Python 3.10+, Node.js 22.12+ and Bash 4.3+ for the local launcher. After make setup, run make demo and open the tracker. The demo needs no API key: portfolio records are read-only, and synthetic manual research uses separate storage. It never overwrites your holdings.

Record trades with fees and fractional shares, preview every balance change, and confirm once. Import standard CSV with duplicate detection, reverse mistakes, back up the ledger, and keep a thesis with a review date for each security. Native CNY/HKD/USD cash, valuation guards and flow-adjusted drawdown make the accounting assumptions visible. AI prepares proposals for your review.

Ledger guide / 账本使用说明 · Release notes

See accounting conventions, planned work, changes and deployment scope.

Stock research workbench / 股票研究工作台

After setup, run make lab and open the workbench. Start a manual experiment on synthetic data without credentials or account setup. Import stock daily bars from CSV, local cache or a market provider; compare momentum, mean-reversion and volatility methods against a baseline, with time-ordered splits, label-boundary purging, costs, curves and reproducible records.

Agent mode plans experiments, reviews validation evidence within a budget, freezes a candidate and evaluates its final test result. Connect an API or an existing official Codex/Claude Code login. Persistent jobs support cancellation, retries, schedules and JSON export; CLI, HTTP, MCP and Python registration support extensions. Connection checks do not guarantee model access or subscription quota.

研究从问题、数据与模板开始，支持手动实验和有次数、时间上限的 Agent 自动研究； 历史记录可以比较、取消、重试和导出。当前聚焦股票日线方法学习与验证，不包含期权回测或实盘下单。

Research data is stored locally. Remote model connections receive research context; exports are not automatically redacted. 研究数据保存在本机；远程模型会收到研究上下文，导出不会自动脱敏。

Research guide · 中文研究指南 · Extension interfaces

Quick Start / 快速开始

git clone https://github.com/Stepuuu/portfolio-daily-tracker.git
cd portfolio-daily-tracker
python3 -m venv .venv
source .venv/bin/activate
make setup
make lab   # research without account setup; open /lab
# make demo   # fictional portfolio and synthetic manual research
# make start  # your configured portfolio plus research
# Open http://localhost:3000

Docker is optional. The commands above run directly on a host or inside an existing container. For users with a Docker host:

docker compose up -d

Messaging integrations / 消息接入

Portfolio tracking, the dashboard and stock research run without OpenClaw. The current Feishu report sender and card bridge use OpenClaw; see the setup guide if you want this integration.

组合追踪、网页和股票研究可独立运行。当前飞书日报发送器与卡片桥接使用 OpenClaw， 需要消息接入时再按配置指南启用。

See full documentation: English · 中文

Contributing / 贡献

We welcome issues, feature requests, and pull requests! If you find a bug, have an idea, or want to improve the code, please open an issue or submit a pull request.

欢迎提交 Issue、功能建议和 Pull Request！ 如果你发现了 Bug、有改进想法或希望贡献代码，请在 Issues 中反馈，或直接提交 PR。

Citation / 引用

If this project helps your research or workflow, please cite it:

如果本项目对你的研究或工作有帮助，欢迎引用：

@misc{portfolio-daily-tracker,
  author       = {Stepuuu},
  title        = {Portfolio Daily Tracker — Self-hosted Investment Portfolio Tracking \& AI Trading Assistant},
  year         = {2026},
  publisher    = {GitHub},
  howpublished = {\url{https://github.com/Stepuuu/portfolio-daily-tracker}},
}
↑↓ navigate openesc close