title	app_file	sdk	sdk_version
KnowSphere – Agentic AI Research Assistant	deep_research.py	gradio	5.33.1

🧠 KnowSphere – Autonomous Deep Research Assistant Using Agentic AI

KnowSphere is an AI-powered research assistant that uses a multi-agent system to generate deep, high-quality research reports on any topic you input. It automates the entire process—from planning, searching, and synthesizing results, to writing detailed markdown reports and emailing them to a recipient—all in a few seconds.

---

🚀 Project Highlights

• Agent-based architecture using simulated Runner, trace, and OpenAI-powered agents.
• Multi-phase pipeline: planning → web search → report writing → email delivery.
• Uses gradio for an interactive and beautifully themed UI interface.
• Custom agents.py enables modular and reusable agent behaviors.
• Deploy-ready on Hugging Face Spaces or local environments.

---

🧠 Research Pipeline Breakdown

1. 🔮 Planning

• The PlannerAgent suggests 3 search queries for a given topic.
• Powered by OpenAI models (gpt-4o-mini).

2. 🌐 Web Searching

• The SearchAgent simulates retrieval of real-time data and summarizes findings concisely.
• Multiple search queries are processed in parallel using asyncio.

3. 📝 Report Writing

• The WriterAgent converts search summaries into a structured markdown report.
• Includes follow-up questions and a short summary.

4. 📧 Email Delivery

• The EmailAgent uses the SendGrid API to send the report via email in HTML format.

---

🖼️ User Interface (Gradio)

Component	Description
Textbox	Input your research topic
Run Button	Triggers the agent pipeline
Markdown Output	Displays the generated markdown report

Custom UI theming with:

• 🎨 Gen-Z pastel styling
• ✨ Animated trace ID log links
• 💡 Bright dark-mode-friendly output panels

---

🖼️ Interface Preview

Fig: The panel accepts your research query, and displays the generated markdown report in a styled dark-mode UI.

---

📂 File Overview

Filename	Description
deep_research.py	Main Gradio app with styled UI and async flow
research_manager.py	Orchestrates the full research workflow
agents.py	Core logic: trace, Runner, Agent, decorators (you create this)
search_agent.py	Agent to simulate real-world search summarization
planner_agent.py	Agent that generates 3 query plans
writer_agent.py	Agent that formats and structures report content
email_agent.py	Sends the generated report to the specified recipient

---

📬 Example Workflow

1. User inputs: "Role of AI in sustainable agriculture"
2. Agents plan relevant queries
3. Agents simulate summarizing real web search results
4. Writer agent formats a detailed markdown report
5. The report is emailed to the predefined address

---

🛠 Setup & Run

1. Install dependencies

pip install -r requirements.txt

2. Set your environment variables

Create a .env file with:

SENDGRID_API_KEY=your-sendgrid-key

3. Run the app locally

python deep_research.py

4. Or deploy on Hugging Face:

gradio deploy

---

📌 Requirements

• Python 3.10+
• Gradio ≥ 4.x
• OpenAI API (for GPT agents, if extended)
• SendGrid (for email sending)

---

💡 Future Enhancements

• Real-time web scraping agent integration
• Dynamic topic summarization with graph plots
• Export reports as .pdf or .docx
• Voice input and TTS report playback

---

Crafted with ❤️ by an autonomous agent crew to make research effortless.