Try HHAI
with AI

Chat directly with Grok-4.5. Ask about open data, APIs, RAG pipelines, or Hong Kong urban systems. Drop the HHAI_RAG_SKILL.md file into the message box to give the agent the full resource catalog as context — then ask it to build agents, write code, or critique ideas.

An API (Application Programming Interface) is a live data endpoint you can call from code or an agent. Instead of downloading a file, you send a request and get structured data back in real time. For example: GET /buildings?district=kowloon returns current building records from a government portal.

How to use it: Pick an API-Ready resource, open its URL in a browser or Postman, read the documentation, then write code (Python, JS, curl) to call it -- or let the API Query Builder agent write that code for you.

RAG (Retrieval-Augmented Generation) is a technique where an AI agent searches a document corpus before generating a response -- grounding its answers in real sources rather than just its training data. You give the agent a collection of PDFs, repos, or web pages; it retrieves the relevant passages; then it answers your question with citations.

How to use it: RAG-Ready resources are documents pre-checked for quality and relevance. Load them into a vector database (Pinecone, Chroma, Weaviate) or paste their text into the Resource Explainer agent to query them directly.

An AI agent is a language model given tools, goals, and the ability to take sequential actions. Instead of just answering a question, an agent can call an API, retrieve a document, write and run code, or chain multiple steps together to accomplish a task. You give it a goal; it works out the steps.

Skill-Ready resources are repos and notebooks that can be wrapped as agent tools -- the agent calls them as functions. This entire platform is designed around making every resource agent-accessible in some form.

Two files — one for any AI assistant, one for your code editor. Paste HHAI_RAG_SKILL.md into Claude, GPT, LangChain or RAGFlow. Drop .cursorrules into your project root for Cursor or Copilot.

Two ready-to-use files — one for any AI assistant, one for your code editor. Download, drop in, and your tool immediately knows the full HHAI catalog, every live HK API endpoint, and how to ingest any resource into a RAG pipeline. No account, no installation, no API key required.

Six live Hong Kong government endpoints callable from Python, JavaScript, curl, or any agent tool. No registration required. Click any row to expand its code example.

python
import requests
r = requests.get("https://www.bd.gov.hk/en/resources/others/opendata/BUILDING_DETAILS.json")
data = r.json()
# Returns: building ID, address, district, year completed, GFA, usage class

python
import requests
query = "Causeway Bay"
r = requests.get(f"https://geodata.gov.hk/gs/api/v1.0.0/locationSearch?q={query}")
print(r.json())
# Returns: place name, x/y coordinates (HK1980 grid), address string

python
import requests
# Get car park list
parks = requests.get("https://static.data.gov.hk/td/parking-vacancy/dataspec/car_park_list.csv").text
# Get live vacancy (replace CP001 with actual park ID)
vacancy = requests.get("https://api.data.gov.hk/v1/carpark-info-vacancy?carparkIds=CP001&lang=en")
print(vacancy.json())
# Returns: car park name, vacancy count, total spaces, timestamp

python
import requests
r = requests.get("https://api.data.gov.hk/v1/historical-air-quality?lang=en")
data = r.json()
# Returns: station name, AQHI value 1-10+, health risk category, timestamp
# Stations: Central, Causeway Bay, Mong Kok, Tuen Mun, Tung Chung...

bash
# General filter API pattern:
curl "https://api.data.gov.hk/v2/filter?q={\"resource\":\"RESOURCE_ID\",\"section\":1,\"format\":\"json\"}"
# Replace RESOURCE_ID with any dataset ID from data.gov.hk

bash
# Browse 800+ datasets at:
# https://data.gov.hk/en/
# Each dataset page has an API tab with its endpoint URL
# Use the filter API (C05) to query any dataset programmatically

All endpoints are read-only, no API key required. Right-click any code block and select "Copy" to get the full snippet. For more datasets visit data.gov.hk.

PDFs and repositories pre-checked for quality, relevance, and ingestion suitability. Download, chunk, embed, and query. Each entry includes recommended chunking strategy.

Four routes from beginner to advanced. Pick the one that matches your tools.

git clone https://github.com/infiniflow/ragflow && cd ragflow
docker compose up -d
# Open http://localhost → upload HHAI_RAG_SKILL.md
# Create agent → connect knowledge base → ask questions

from langchain.document_loaders import TextLoader
from langchain.text_splitter import MarkdownTextSplitter
from langchain.vectorstores import Chroma
from langchain.embeddings import OpenAIEmbeddings

docs = TextLoader("HHAI_RAG_SKILL.md").load()
chunks = MarkdownTextSplitter(chunk_size=800).split_documents(docs)
vectorstore = Chroma.from_documents(chunks, OpenAIEmbeddings())
retriever = vectorstore.as_retriever(search_kwargs={"k": 4})

results = retriever.get_relevant_documents(
    "How do I get air quality data for Hong Kong?"
)

# .cursorrules
Refer to the HHAI RAG catalog for HK open-data resources.
For HK data questions, first check C01-C06 APIs.
For RAG pipeline questions, recommend B10 (IBM) or B11 (RAGFlow).
Catalog: HHAI_RAG_SKILL.md in project root.