TrendRadar AI: Start With a Smaller News Brief
If keeping up with your industry means checking the same sites each morning, TrendRadar can gather and filter their headlines into a report. Start with a few RSS feeds and clear keyword rules, then add AI features when you know what you want them to do.
sansan0’s TrendRadar is an open-source news and trend monitor. This article covers that project, including a small test of its English keyword matcher.
The useful first goal is a brief you would actually read: relevant sources, a manageable topic and links worth opening.
Key Takeaways
- TrendRadar combines trending lists and your chosen RSS feeds, with optional AI filtering, translation and analysis.
- Plain keywords match parts of words. In our small test,
AIalso matchedrailandtraining.- The starter rules below use whole-word matches and one exclusion. They were checked with nine invented English headlines.
- Collection, AI analysis and notification delivery need separate setup. Our trial tested the keyword module, rather than the complete app.
How does TrendRadar change your daily reading?
TrendRadar changes your daily reading by putting trending lists and the RSS feeds you choose into one filtered report, so both the sources and the topics become decisions you make instead of defaults.
You choose the sources. RSS is a feed of updates published by a site. TrendRadar can bring those feeds into the same workflow as its trending lists. For an English-language work brief, choosing suitable feeds matters more than importing every available platform.
The RSS configuration guide explains how to add a feed and limit how old its articles can be.
You choose what gets through. Keyword rules can include topics, require words and exclude unwanted terms. A report about AI tools for work might follow AI, OpenAI and Claude, then exclude a promotion marker you keep seeing.
You choose how to receive it. The project produces HTML reports and supports delivery through services such as Telegram, email and Slack. A local file and a message arriving on your phone are separate outcomes; each needs checking.
AI can handle a more flexible request later. With AI smart filtering, you describe your interests in ordinary language. The model scores news against them. That may suit a topic with several ways of saying the same thing.
I would first check whether the chosen sources contain useful stories. A clever filter has little to work with when the input is wrong.
Can a keyword rule catch the wrong story?
A plain keyword rule can catch the wrong story because it matches parts of words, and in our test a rule for AI also matched headlines about rail services and team training.
TrendRadar’s keyword configuration supports both plain text and regular expressions. A regular expression is a matching pattern; the \b markers below mean a word boundary.
We ran the project’s unmodified keyword module against nine invented English headlines. Both rule files used the same input.
| Constructed headline | Plain topic words | Whole-word rules + exclusion |
|---|---|---|
| AI tools for meeting notes | Keep | Keep |
| Rail services extend weekend hours | Keep | Drop |
| Team training moves to Friday | Keep | Drop |
| New OpenAI model for speech | Keep | Keep |
| Claude adds a document workflow | Keep | Keep |
| Weekend football fixtures | Drop | Drop |
| Paid AI course: coupon offer | Keep | Drop |
| AI-powered search reaches new users | Keep | Keep |
| Daily rainfall report | Keep | Drop |
The plain file contained three lines: AI, OpenAI and Claude. It kept eight headlines because matching ignores case and accepts those letters inside longer words.
The revised file kept four. Copy this into config/frequency_words.txt in a test configuration:
[GLOBAL_FILTER]
coupon
[WORD_GROUPS]
/\bAI\b/ => AI
/\bOpenAI\b/ => OpenAI
/\bClaude\b/ => Claude
The three patterns belong to one group. A title can match any of them. The global filter excludes titles containing coupon.
Use that exclusion only if it fits your reading goal. A coupon may be relevant to a retail report. Likewise, these whole-word patterns will miss related stories that never name AI, OpenAI or Claude.
This is why I would inspect the first results before adding more rules. You need to see both what was kept and what disappeared.
Test scope: The check used Python 3.12 and the official v6.11.0 keyword module, retrieved on October 29, 2026. Network access was blocked during the isolated check. These results show matching behavior on a constructed sample; they do not measure real-news accuracy or time saved.
How small should the first TrendRadar setup be?
The first setup should cover one reading task and a handful of feeds, because a narrow question makes it easy to reject distracting stories.
Choose one reading task. For example: “Find changes to AI document tools that could affect our weekly workflow.”
That gives you a reason to choose sources and reject distracting stories. The same idea helps when writing a brief with a clear task and checks.
Then choose a small number of feeds that regularly publish relevant material. You can expand the list after reading the first report.
For an existing installation, check these settings in config/config.yaml before trying the keyword-only route:
| Setting | Starting value | Purpose |
|---|---|---|
filter.method |
keyword |
Use the text rules |
ai_analysis.enabled |
false |
Leave generated analysis off |
ai_translation.enabled |
false |
Leave generated translation off |
schedule.enabled |
false |
Keep per-period strategies from overriding this initial configuration |
These keys are present in the project’s English configuration template. Edit the corresponding values in your complete configuration; the table is not a replacement configuration file.
If you are starting from scratch, follow the official local deployment guide. The project requires Python 3.12 or later and documents local, Docker and GitHub Actions options. Initial installation and configuration may need technical help.
Our trial did not run that complete installation, fetch RSS, generate a TrendRadar report or deliver a notification. After setting up your own instance, verify those stages before relying on a scheduled brief.
When do AI and MCP become useful in TrendRadar?
Keyword rules work while you can name the topic in one or two words, and AI filtering, translation or MCP become worth the setup once relevance depends on context you cannot write as a term.
AI filtering lets you describe that context, such as a particular industry or a kind of product change. Translation and analysis can then help you read or summarize the selected material.
The project FAQ says the core collection, keyword filtering and push functions can operate without AI. Model-backed features add provider configuration and potential usage charges. The open-source license does not make every hosting or model option free.
MCP stands for Model Context Protocol. It lets a compatible assistant use tools to work with collected news. One possible task is asking it to find stories about a topic in the stored data. It is an extra capability to explore after the collection is useful.
We did not test AI relevance scores, translations, summaries or MCP responses in this trial.
Who is TrendRadar for?
TrendRadar suits a writer, an analyst or anyone who follows one recurring subject and is willing to maintain a small tool; a plain RSS reader is enough if you only want to skim a few sites.
A writer following a recurring subject. A focused reading list can supply leads for articles. You still need to open the sources and check whether a lead deserves a story.
Someone preparing an industry update. A repeatable source list and explicit rules make it easier to explain what the brief covers. Keep notes on important additions and exclusions.
A person comfortable maintaining a small tool. This project gives you configuration choices, along with installation and upkeep. If you simply want to follow a few sites, an ordinary RSS reader may be enough.
Self-hosting also describes where the program and data run. Your selected feeds, trending-data service, notification service and AI provider can still involve external requests. Check those choices before adding confidential material.
What is the first step with TrendRadar?
The first step is to write down the one question your next brief should answer, then pick three topic terms and review a small batch of headlines before you change any other setting.
Pick three topic terms and review a small batch of headlines, including some you expect to reject.Keep a note of one useful result and one poor match. A shared worklog can help if more than one person or assistant maintains the rules.
Make one change, then inspect the next results. A brief earns its place in your day when its contents help you make the decision you started with.
About the author: JH-A writes practical guides to AI tools, digital workflows and design.
This post was drafted with the help of AI writing tools, based on the documented keyword-module trial and official project documentation. The cover is an AI-assisted editorial illustration, not a screenshot of TrendRadar.
Frequently Asked Questions
Does TrendRadar need an AI model to work?
No. The project FAQ says collection, keyword filtering and push notifications work without AI; model-backed features add provider configuration and possible usage charges on top.
Why did a rule for AI match headlines about rail services?
Because plain keyword matching matches parts of words. The starter rules in this article use whole-word matching plus one exclusion, checked against nine invented English headlines.
Is TrendRadar free?
The project is open source, but that does not make every hosting or model option free. Self-hosting and any model usage can carry their own costs.
How many feeds should I start with?
A small number that regularly publish material for one reading task. You can expand the list after you have read the first report and seen which sources earned their place.
Comments
Post a Comment