Update README.md

add demo video links and tags

README.md

@@ -9,6 +9,9 @@ app_file: app.py
 pinned: false
 license: apache-2.0
 short_description: Egg Tart - 蛋撻仔 🍮 Agent of Hong Kong Data
+tags:
+   - mcp-server-track
+   - agent-demo-track
 ---
 
 # 🥚 Egg Tart - 蛋撻仔 🍮: Your Hong Kong Data Agent
@@ -17,6 +20,40 @@ Egg Tart - 蛋撻仔 🍮 is a friendly, LLM-powered assistant that answers your
 
 ---
 
+## 🎯 Agents-MCP-Hackathon 2025 Submission
+
+**Tracks:**  
+- Track 1: MCP Tool / Server  
+- Track 3: Agentic Demo Showcase
+
+**Tags:**  
+- `mcp-server-track`
+- `agent-demo-track`
+
+**Demo Video:**  
+- Track 1 Demostration: Video recording of Egg Tart Agent - MCP server in action using n8n MCP Client >> [Track 1 Demo Video](https://youtu.be/r6gDzpL0kIM?si=B-UmGgEuKxs9Pp0O) <<
+- Track 3 Demostration: Overview Video of Egg Tart Agent explaining the usage / purpose of the application  >> [Track 3 Demo Video](https://youtu.be/KdY78nZUcDg?si=nT7E5vsMwSkzXqv2) <<
+
+### Track 1: MCP Tool / Server
+
+Egg Tart provides an MCP-compatible server interface powered by gradio UI, enabling LLM agents to answer questions about Hong Kong data.
+
+- Powered by the smolagents framework and Gradio UI.
+- Agent execution is exposed as an MCP Tool / Server via Gradio (`mcp_server=True`).
+- The exposed function `ask` is well-documented for easy integration.
+- The demo shows how the n8n chat agent calls the app as a tool.
+
+### Track 3: Agentic Demo Showcase
+
+Egg Tart demonstrates a conversational agent that automates the discovery, understanding, and retrieval of open data from data.gov.hk.
+
+- Users ask questions in natural language; the agent handles the rest.
+- The agent leverages semantic search, dataset exploration, content fetching, and PDF-to-markdown parsing.
+- Powered by smolagents and Gradio.
+- The demo highlights how LLM agents and smolagents streamline access to reliable, documented, and up-to-date public datasets, with full citations and context.
+
+---
+
 ## ❓ Why Egg Tart?
 
 Accessing open government data from data.gov.hk can be challenging. Datasets are scattered across different endpoints and use inconsistent data dictionaries. For example, to find the weather forecast for Hong Kong, you must:
@@ -82,41 +119,6 @@ Egg Tart automates this entire process using smolagents and custom tools, making
 
 ---
 
-## 🎯 Hackathon Submission
-
-**Tracks:**  
-- Track 1: MCP Tool / Server  
-- Track 3: Agentic Demo Showcase
-
-**Tags:**  
-- `mcp-server-track`
-- `agent-demo-track`
-
-**Demo Video:**  
-- [PLACEHOLDER: Insert demo video link here]
-
-### Track 1: MCP Tool / Server
-
-Egg Tart provides an MCP-compatible server interface powered by gradio UI, enabling LLM agents to answer questions about Hong Kong data.
-
-- Powered by the smolagents framework and Gradio UI.
-- Agent execution is exposed as an MCP Tool / Server via Gradio (`mcp_server=True`).
-- The exposed function `ask` is well-documented for easy integration.
-- The agent uses semantic search, dataset exploration, and content retrieval tools.
-- If parameters are required, the agent parses the data dictionary (PDF) using docling.
-- The MCP server returns results in a standardized format for easy integration.
-- The demo shows how the n8n chat agent calls the app as a tool.
-
-### Track 3: Agentic Demo Showcase
-
-Egg Tart demonstrates a conversational agent that automates the discovery, understanding, and retrieval of open data from data.gov.hk.
-
-- Users ask questions in natural language; the agent handles the rest.
-- The agent leverages semantic search, dataset exploration, content fetching, and PDF-to-markdown parsing.
-- Powered by smolagents and Gradio.
-- The demo highlights how LLM agents and smolagents streamline access to reliable, documented, and up-to-date public datasets, with full citations and context.
-
----
 
 ## 💡 How to Use