Nathan Wrigley of the Jukebox Podcast (WP Tavern) talks with developer and long‑time WordPress contributor Damon Cook about WP Trend Watcher, an open source tool Damon built to help busy WordPress professionals stay on top of the ever‑growing flow of news, core discussion, plugin and theme updates, and community activity.
Who Damon Is
Damon has worked in the WordPress ecosystem for more than a decade, mostly at WordPress‑focused agencies and more recently as a freelancer running Noma Digital. He remains engaged with core development and the wider community and watches releases such as WordPress 7.1 closely.
The Problem WP Trend Watcher Solves
There’s too much content. Between blogs, Slack, Trac, social networks and changelogs, manually scanning everything is time-consuming and inefficient. Damon started WP Trend Watcher for himself: to collect RSS feeds, have them summarized, and then verify those summaries so he can rapidly gauge what matters without blindly trusting AI output.
How It Works (Workflow and Philosophy)
– Source collection: The tool aggregates a curated list of RSS feeds from WordPress news sites, developer blogs, and community sources. Damon is open to adding more sources via contributions.
– Local AI summarization: WP Trend Watcher runs prompt‑based summaries using local LLMs through tools such as LM Studio. Damon experiments with different models (he’s used variants including Qwen and others) depending on what his machine can run—he mentioned having around 18 GB of memory to work with, which constrains model size.
– Prompting for relevance: Summaries are generated with prompts targeted at developers and agency owners, emphasizing factual, developer‑focused takeaways rather than clicky popularity metrics.
– Deduplication and filtering: The system tries to detect duplicate coverage across sources and removes weak or repetitive items so the final list is more useful.
– Human in the loop: Crucially, the AI output is not published as‑is. Damon reviews the summaries, clicks through the original articles, fact‑checks to prevent hallucinations, and adds his own human notes. He considers this step essential both to accuracy and to internalizing the information.
Output and Publication
Damon runs the tool on a roughly weekly cadence (it can be configured to run more often). He compiles the results into a short, numbered list of items—each with a brief AI‑generated summary and a link—followed by his human commentary. He publishes the final briefing to GitHub Pages from a TypeScript codebase hosted on GitHub. He intentionally keeps the publication process semi‑manual because he values the human curation stage.
Why Not Fully Automate?
Damon draws a line between automation and AI assistance. Fully automating the pipeline (AI summarizes and publishes without human review) would be possible technically, but would defeat the main purpose: reliable, contextualized information for developers who need to act on or advise clients about upcoming changes. The human step also helps the content stick—reading, checking and adding notes makes the update more memorable.
Transparency and Open Source
WP Trend Watcher is open source. Damon publishes which model he used and how much time he spent auditing the report so readers can judge the depth of curation for themselves. He encourages forks and contributions so others can adapt the project or add sources.
Technical Notes
– Codebase: TypeScript, available on GitHub (Damon’s handle is colorful‑tones on some platforms).
– Local models: Runs locally via LM Studio or comparable tooling; model choice depends on available RAM and performance tradeoffs.
– Publication: Reports are pushed to GitHub Pages; the manual review step is what makes the pipeline intentionally non‑fully automated.
Accuracy and Limits
Damon reports the summaries are generally accurate when the prompts and model work well. Still, he treats the process like a game of fact‑checking: he checks AI summaries against full articles to guard against hallucinations and to catch important details the summary might miss. The system also filters duplicates and weak posts, but scope creep is a risk if too many sources are added.
Future Possibilities
One natural extension Damon envisions is a trend analysis layer: using the accumulated weekly reports as a corpus, an AI could retrospectively identify patterns—why something progressed from a proposal to core, where discussion clustered, or how market‑level trends shifted over many reports. Over time, these historical receipts could be used to answer deeper “how did we get here?” questions across any industry, not just WordPress.
Who Benefits
WP Trend Watcher is aimed at developers, freelancers, agency owners, and anyone in the WordPress ecosystem who needs to anticipate change and advise clients. It’s particularly useful for people who want to be “one step ahead” of feature releases and community shifts without spending hours manually scanning feeds.
How to Try It
Damon documents setup steps on his site and the GitHub repo: you can fork the TypeScript project, provide RSS sources, run local models via LM Studio or similar, and configure the cadence you prefer. The published weekly briefings are available to review on the project’s GitHub Pages so you can judge the format and depth before installing anything.
Final Thoughts
WP Trend Watcher is a practical example of combining AI with human curation: the tool narrows the noise, AI produces concise summaries, and a human validates and contextualizes the output. Damon intentionally keeps the human step and publishes his process openly so others can adapt or improve it. The project shows how local models and curated sources can make staying current manageable, and it hints at richer longitudinal analysis possibilities as the archive of reports grows.
If you want to explore the project, check the WP Tavern episode notes for links to Damon’s GitHub, his blog post with setup details, and the weekly reports themselves. Damian also shares the reports on social channels for anyone who wants a quicker way to follow what’s happening in WordPress.
