← back to the blog

ScholarSynth: Hiring My First AI Agent

January 26, 2025

It all started during my master's degree — I got hooked on research papers. So many innovations, so many novel ideas. But when you have to read a 20-page research paper, the process is brutal. It goes like this:

1. Open arXiv or Google Scholar.
2. Click through dozens of confusingly titled papers.
3. Download the PDF.
4. Read through a mountain of words that does not make sense (sometimes).
5. Try to summarize it before your brain melts.

And I thought, "There must be an AI for this"? No? Just me? Well, welcome to my world. With the rise of AI agents — autonomous systems that perceive their environment, take actions to achieve goals, and improve their performance over time — ScholarSynth was born.

What Is ScholarSynth?

ScholarSynth is like a super-efficient research intern — except it never gets tired, never asks for a raise, and doesn't text its friends in the middle of a task. This AI agent can:

• Search for research papers like a detective hunting for clues.
• Extract key information so you don't have to read all 30 pages.
• Summarize findings in a way that actually makes sense.
• Format them into blog posts (because why not share the knowledge?)

Automating Research Paper Reading with ScholarSynth

The moment I realized AI could handle the "reading & summarizing" part while I focused on actual learning, I knew I had to build this. Reading papers manually is like assembling IKEA furniture without instructions — you'll eventually get there, but at what cost?

The Dream Team: ScholarSynth's AI Agents

🕵️ Agent 1: ArXiv Paper Detective

This agent is responsible for finding fresh research papers, so I don't have to manually visit arXiv every day. It scans recent publications and selects one I haven't processed yet.

📖 Agent 2: PDF Whisperer

Once we have the paper, we need to extract its main content.

• Reads the PDF like a diligent student.
• Extracts key sections — Abstract, Methodology, Results, Conclusion.
• Prepares the text for summarization.

✍️ Agent 3: The Over-Achieving Summarizer

This agent summarizes the extracted content into a digestible format.

• No unnecessary fluff — just key takeaways.
• Maintains technical accuracy while making it easy to understand.

📰 Agent 4: The Blog Post Magician

Now we have a summary, but let's make it fun & readable. This agent:

• Adds an engaging introduction.
• Formats content like a blog (headings, sections, bullet points).
• Ensures the post is SEO-friendly (because AI is nothing if it's unreadable).

Making ScholarSynth Work Automatically

You ever have a hobby project that only works when you remember to run it? Yeah, that's not sustainable. I wanted ScholarSynth to work on its own, so I set up:

• ✅ GitHub Actions to automate execution.
• ✅ Git commits to track progress.
• ✅ Auto-publishing to a blog

The Cost Breakdown 💰

If you're wondering how much this automation costs, here's the best part:

• Crew AI → Free 🆓
• GitHub Actions → Free 🆓
• OpenAI API → ~$0.03 per research paper (approximate, varies by length)

That's it! ScholarSynth runs at practically zero cost, except for OpenAI processing fees, which are negligible.

Final Thoughts

ScholarSynth started as a simple way to automate research reading, but it turned into something much bigger — a hands-free AI-powered research assistant.

If you've ever struggled to keep up with research papers, ScholarSynth can do the heavy lifting. And if you build something similar, let me know — I'd love to see your take on it!

← back to the blog