WeeklyMark is a local-first, Python-powered engine that scans your Obsidian or Markdown vault to extract tasks, tags, and highlights. It generates a structured weekly report in seconds, removing the need to manually hunt through dozens of notes to organize your review.
💎 THE ELITE ADD-ON: ChangePoints OS
For the practitioners who need data, not just documentation.
While WeeklyMark manages your personal workflow, ChangePoints OS provides the high-level data backbone for serious ESG leads and policy designers.
Stop guessing what drives sustainable behavior. The ChangePoints Notion Vault is a digitized, searchable database of 50 historically proven ESG policy interventions.
- Rigorously Scored: Every intervention, such as Rwanda’s plastic ban or London’s ULEZ, is scored by actual behavioral impact, anomaly, and adoption speed.
- Fully Cited: The database is backed by primary source data from organizations like the World Bank and IMF.
- Ready to Deploy: Features one-click duplication directly into your own Notion workspace.
Launch partners are being sought in the ESG and Behavioral Science space to join the program.
- 70% Commission: For the first 30 days, the standard affiliate split is bumped to 50% for partners.
- Done-for-you Marketing: Partners receive Bento Grid graphics and pre-written copy to simplify promotion.
- Lifetime Access: Launch partners receive the full OS and all future data updates for free.
Apply to be a ChangePoints Partner here
Using Obsidian as a "Second Brain" often turns the weekly review into a manual chore. Users must open every daily note to find specific tags, check open tasks, and synthesize a summary manually.
WeeklyMark automates the discovery phase of your review.
- Automated Extraction: Pulls every task and tag from notes modified within a specified timeframe, such as the last 7 days.
- AI-Powered Synthesis: Utilizes models like GPT-4o-mini to transform raw tasks and tags into a concise executive summary.
- Dual-Mode Flexibility: Offers a CLI for automated workflows and a PyQt6 GUI for a standard desktop experience.
- Smart Date Parsing: Supports multiple frontmatter date formats, including ISO and YYYY-MM-DD.
- Regex-Powered Task Tracking: Detects bullets, checkboxes, and nested tags without disrupting Markdown syntax.
- Local-First: Your vault stays on your machine; tasks are only sent to OpenAI if the summarize flag is enabled.
- Custom Templates: Reporting is fully customizable via Jinja2 templates.
-
Clone and Install:
git clone [https://github.com/Rage47/WeeklyMark.git](https://github.com/Rage47/WeeklyMark.git) pip install -r Requirements.txt
-
Run the GUI (Desktop App):
python __main__.py
-
Generate a Report (CLI mode):
python __main__.py ~/MyVault --summarize --output ./Weekly-Report.md
- Python 3.9+
- PyQt6 (for GUI)
- Jinja2 (for rendering)
- python-frontmatter (for parsing metadata)
- openai (for summaries)
I built WeeklyMark because I love Obsidian, but I absolutely hated my Sunday afternoon "Weekly Review" ritual. Digging through dozens of daily notes just to find scattered checkboxes and tags felt like a massive waste of time. I needed a local-first engine to do the heavy lifting so I could actually focus on planning my week.
I decided to open-source the core Python logic because I believe the Markdown and PKM (Personal Knowledge Management) community thrives on shared tools. You are completely free to clone this repo, tweak the Jinja templates, and run it locally forever.
Want to skip the terminal?
If you don't want to mess with pip install, Python virtual environments, or executing scripts every Sunday, I have compiled WeeklyMark into native, 1-click desktop applications for macOS and Windows.
You can grab the plug-and-play desktop app (and support my late-night coding sessions) over on Gumroad:
MIT License.
Developed by Alex Rotar. Building tools for the intersection of Behavioral Economics and ESG.