🧠 ThinkDigest – Research Topic Summarizer Agent (Mastra)

ThinkDigest is an intelligent agent-powered assistant that summarizes recent research on any topic using LLMs. It is built using Mastra, a TypeScript framework for building LLM-powered agents and workflows.

---

✨ Features

• 🔍 Extracts a relevant topic from a user query
• 📄 Fetches 5 recent research papers using dynamic topic resolution
• 🧠 Summarizes key trends and ideas across those papers
• 🗣️ Returns a human-readable digest of recent research
• 🧪 Works locally via Mastra dev interface (http://localhost:4111)

---

🛠 Tech Stack

• Mastra for agent orchestration
• Gemini 2.0 Flash (via Google AI API) as the LLM backend
• TypeScript for implementation
• Zod for schema validation

---

📦 Installation

1. Clone the Repo

git clone https://github.com/lakshani09dl/thinkdigest.git
cd thinkdigest

2. Install Dependencies

npm install

3. Setup TypeScript

Make sure you have tsconfig.json like:

{
  "compilerOptions": {
    "target": "ES2022",
    "module": "ES2022",
    "moduleResolution": "bundler",
    "esModuleInterop": true,
    "forceConsistentCasingInFileNames": true,
    "strict": true,
    "skipLibCheck": true,
    "noEmit": true,
    "outDir": "dist"
  },
  "include": [
    "src/**/*"
  ]
}

4. Run Locally

npm run dev

Mastra will start the agent runner at: 👉 http://localhost:4111

UI Preview