In a recent Jukebox Podcast conversation, Damon Cook explained why he built WP Trend Watcher and how it helps WordPress professionals keep up with an increasingly noisy ecosystem. Damon, a developer with more than a decade working around WordPress and now running his freelance shop Noma Digital, set out to solve a simple but time-consuming problem: staying informed without spending hours chasing links.
What WP Trend Watcher does
WP Trend Watcher is an open source tool that aggregates RSS feeds from WordPress-related sources, summarizes the incoming items with local AI models, and surfaces a concise weekly briefing aimed at developers, agency owners, and freelancers. The goal is not to replace reading, but to reduce the “dead time” spent finding signals in the noise—so you can quickly identify items worth clicking through and thinking about.
Why Damon built it
Damon started the project for personal reasons: he wants to be aware of new features, proposals, and community changes so he can anticipate the needs of clients and respond knowledgeably when things land in Core or in popular plugins and themes. He also wanted to experiment with local AI models while keeping a human in the loop to verify summaries and avoid hallucinations. The result is a workflow that combines automation for discovery and AI-generated summaries with deliberate human review and commentary.
How it works (workflow and tech)
– Sources: The tool gathers a set of curated RSS feeds from around the WordPress ecosystem. Users can suggest sources or fork the project to add their own.
– Local inference: WP Trend Watcher is a TypeScript codebase you run locally. Damon uses LM Studio on his Mac to host smaller, local LLMs that fit his machine (he notes system memory limits when choosing models). Each run pulls new items since the last run, feeds article text into a model, and asks for a summary tailored to developers and agency owners.
– Deduplication and filtering: The prompts and logic are designed to detect duplicates across publications, drop weak or superfluous pieces, and focus on factual, developer-relevant content rather than simply what’s trending.
– Human-in-the-loop: After the AI generates summaries, Damon reviews the list, clicks through high-interest items, fact-checks, and writes a short human summary or judgment. That step is deliberate—he finds reading and annotating helps the information stick and prevents blind trust in AI outputs.
– Publishing: Once curated, Damon publishes the weekly briefing to a public GitHub Pages site. He runs the process manually on a schedule that suits him (weekly by default), then commits the results. The human review step intentionally prevents full automation for now.
Design choices and values
Damon built in transparency: each report notes the model used and the amount of time he spent reviewing items. He doesn’t shy away from admitting that using AI makes him feel a bit “cheap,” but publishing the process and inviting others to fork the repo makes the work auditable. That transparency lets readers decide whether the level of review is sufficient.
The tool favors accuracy and usefulness over clickbait. It’s tuned to surface items that matter to professionals (feature proposals, Core changes, plugin developments) rather than simply the most popular posts. It also attempts to keep outputs deterministic and plans to add clearer footnotes and more consistent formatting so readers can see exactly which sources informed each summary.
Technical constraints and experimentation
Damon emphasizes that running inference locally shaped many decisions. Model size, performance, and the constant arrival of new models change the outputs week to week, so he experiments with different local models to see how summarization quality varies. The project is intentionally open source to allow others to swap models, add feeds, or adapt the system to other domains.
Bigger picture and future possibilities
While WP Trend Watcher currently targets WordPress, the architecture is domain-agnostic. Over time, keeping a serialized corpus of weekly reports could enable higher-level trend analysis—asking which topics have gained or lost attention, why features moved into Core, or how community conversations evolved. Damon sees the current outputs as scaffolding for future trend-detection layers that could make longitudinal inferences across reports.
Practical takeaway
If you want to try WP Trend Watcher: the project is available on GitHub as open source; it’s a TypeScript project you run locally and pairs with local LLMs via tools like LM Studio. Damon publishes weekly briefings, documents the model used, and links to the source articles so readers can verify claims. WP Tavern’s show notes for the podcast episode include links to the repository, the published briefings, and Damon’s blog post explaining how to get started.
Final note
WP Trend Watcher is a pragmatic experiment in mixing human judgment with AI-powered curation. It reduces the time spent hunting for relevant WordPress news while preserving the human step that validates, contextualizes, and remembers what matters. Damon invites contributions and feedback—if you’re curious about running your own version or adapting the approach to another field, the code and weekly outputs are public and intended to be a starting point, not a finished product.
