I Built an AI Research Agent to Cure My "Doomscrolling" Addiction

The Problem: AI News is Noise

Every morning, I faced the same problem. There are 50 new AI tools released daily, 10 new models on HuggingFace, and endless hype on X/Twitter.

I was wasting hours "doomscrolling" just to find the 2 or 3 updates that actually mattered to my work.

I didn't need more news. I needed a Chief of Staff to read everything for me, filter out the garbage, and only show me the signal.

So, I built one.

The Solution: An Autonomous "News Editor"

In this tutorial, I’ll show you how I built a Personal AI News Agent using n8n, OpenAI, and Tavily.

It works while I sleep:

**Reads **the raw RSS feeds from major tech sites.

**Judges **every headline (acting as a strict "Senior Editor").

**Researches **the winners using Tavily (to verify facts).

**Delivers **a curated morning briefing to my email.

The Stack

Orchestrator: n8n (Local or Cloud).

The Brain (Filter): OpenAI gpt-4o-mini (Cheap and fast).

The Researcher: Tavily AI (Essential for fetching live context).

Source: RSS Feeds (e.g., TechCrunch, Verge).

Step 1: The "Firehose" (RSS Ingestion)

The workflow starts with a Schedule Trigger set for 8:00 AM. It pulls the latest articles using the RSS Read Node.

At this stage, we have everything—rumors, minor updates, and noise.

Step 2: The "Senior Editor" (OpenAI Filtering)

This is the most critical part. I didn't just ask AI to "summarize." I used a Loop Node to process each headline individually and gave OpenAI a specific persona:

_System Prompt: "Analyze this news item:

Title: {{ $json.title }}

Summary: {{ $json.contentSnippet || $json.content }}

YOUR ROLE:

You are a Senior Tech Editor curating a daily briefing. Your goal is to identify useful, relevant news for AI Engineers.

SCORING GUIDELINES (0-10):

• 0-3: Irrelevant, gossip, or low-quality clickbait.


• 4-5: Average news. Minor updates or generic articles.


• 6-7 (PASSING): Solid, useful news. Good tutorials, interesting tool releases, or standard industry updates.


• 8-10 (EXCELLENT): Major breakthroughs, aquistitions, critical security alerts, or high-impact releases (e.g., GPT-5, new SOTA model).

INSTRUCTIONS:

1. Rate strictly but fairly.


2. If it is useful to a professional, give it at least a 6.


3. Return ONLY a JSON object.

OUTPUT FORMAT:

{

"score": ,

"title": ,

"reason": ""

}_

Step 3: The Gatekeeper (If Node)

I added an If Node that acts as a gate.

_Score < 7: Discard immediately.

Score >= 7: Proceed to research._

This simple logic reduced my reading list from ~50 articles to just the top 5.

Step 4: The Deep Dive (Tavily AI)

For the winning articles, I didn't want just the RSS blurb. I used Tavily AI to go out and "read" the full context of the story.

I set Tavily's include_answer parameter to "Advanced." This generates a high-quality, synthesized summary of the topic based on multiple sources, not just the original article.

Step 5: The Briefing (Email)

Finally, an Aggregate Node collects all the "Winners" and formats them into a clean HTML email, sent via Gmail.

Watch the Build (Step-by-Step)

I recorded the entire process, including the exact Prompt and JSON logic I used. You can follow along here:

Why This Matters

By building this agent, I saved myself ~5 hours a week of mindless scrolling. The agent does the boring work of filtering; I just read the high-signal results.

Next Steps: In my next post, I’ll share how I used Google NotebookLM to "stress test" this agent.

Let me know in the comments: How are you handling the information overload right now?