Use Huwise APIs Without Writing a Single Line of Code with Claude Code

You manage a data portal. You know your data, your users, your workflows. But when someone mentions “API calls” or “REST endpoints,” you feel like you’ve crossed into developer territory.

What if you could talk to your Huwise APIs the way you talk to a colleague? That’s exactly what Claude Code makes possible. No programming experience required.

This guide walks you through setting up Claude Code to work with Huwise APIs, running your first real task, and keeping costs under control.

What Is Claude Code?

Claude Code is a command-line assistant made by Anthropic. You type what you want in plain language, and it figures out the technical details — including making API calls, reading responses, and chaining multiple steps together.

Think of it as having a developer sitting next to you who never gets tired of your questions.

You can run it:

• In your terminal (Mac, Windows, Linux)
• In VS Code or JetBrains as an extension
• On the web at claude.ai/code

To install it, open a terminal and run:

npm install -g @anthropic-ai/claude-code

Then launch it by typing claude in your terminal.

Step 1: Get Your API Key

Before Claude Code can talk to your Huwise portal, you need an API key.

1. Log into Huwise 
2. Generate your Huwise API key.

Your API key looks something like: 8a4e52f...b3d1. Keep it private — it grants access to your portal’s data.

Important: Once shared with Claude, you API key is at risk, you can delete it and use a new one for each project.

Step 2: Give Claude Code the Context It Needs

Claude Code works best when you give it clear context upfront. The first time you use it with Huwise, start your conversation like this:

I need to interact with my Huwise data portal APIs.

Here is my configuration:

• Portal domain: https://myportal.huwise.com
• API key: YOUR_API_KEY

The API documentation is available at:

• Explore API: https://developer.huwise.com/apis/explore/v2.1/index.html
• Automation API: https://developer.huwise.com/apis/automation/v1.0/index.html
• Portal API: https://developer.huwise.com/apis/portal/v1.0/index.html

All API requests should include the header: Authorization: Apikey YOUR_API_KEY

That’s it. Claude Code now knows how to authenticate and where to find documentation if it needs to look something up.

Important: Claude Code does not store your API key between sessions. You’ll need to provide it each time you start a new conversation, or you can set it as an environment variable (more on that later in the optimization section).

Step 3: Your First Task — List Your Published Datasets

Let’s start simple. Type this:

Show me the list of all datasets published on my portal

Claude Code will call the Explore API behind the scenes:

GET <https://myportal.huwise.com/api/explore/v2.1/catalog/datasets>

And return something readable like:

No need to know the endpoint, the pagination logic, or the response format. You asked, you got your answer.

A Real-World Use Case: Monthly Data Quality Check

Let’s walk through a scenario that many portal managers face: checking that your key datasets are up to date and contain valid data.

Ask Claude Code to Run Your Check

For each of my published datasets, tell me:

1. When it was last updated
2. How many records it has
3. Whether any text fields contain empty values

Flag any dataset that hasn’t been updated in the last 30 days

Claude Code will:

1. Fetch your dataset catalog using the Explore API
2. For each dataset, query its metadata and record count
3. Run targeted queries to detect empty fields
4. Present a clear summary with flagged issues

The Result

You get a report like this:

DATA QUALITY CHECK — March 31, 2026

✅ City Budget 2025 — 1,240 records — Updated 16 days ago
   No empty fields detected.

⚠️  Public Events Calendar — 412 records — Updated 31 days ago
   WARNING: Not updated in 30+ days.
   3 records have empty "location" field.

✅ Air Quality Sensors — 58,302 records — Updated 1 day ago
   No empty fields detected.

This would have required dozens of individual API calls. You did it with one sentence.

More Things You Can Ask

Once you’re comfortable, the range of tasks is wide. Here are practical examples for each API:

Explore API (Query Published Data)

• “How many records were added to the ‘traffic-counts’ dataset this month?”
• “Show me the top 10 cities by population from the ‘census-data’ dataset”
• “Export the ‘building-permits’ dataset as a CSV file”
• “What are the available values for the ‘category’ field in ‘public-services’?”

Automation API (Manage Your Portal)

• “Create a new dataset called ‘Q1 Survey Results’ and upload this CSV file”
• “Update the metadata description of the ‘parking-lots’ dataset”
• “Publish the ‘annual-report-2026’ dataset”
• “Schedule the ‘weather-stations’ dataset to refresh every day at 6 AM”
• “Add user jean.dupont@city.gov with editor permissions to the ‘budget’ dataset”

Portal API (Manage Content)

• “List all the pages on my portal”
• “Show me the current navigation menu structure”
• “What forms are available and how many submissions does each have?”

Optimize Your Token Consumption

Claude Code uses tokens every time you interact with it. Tokens cost money, so here are practical ways to keep your usage efficient.

1. Set Up a CLAUDE.md File

Create a file named CLAUDE.md in your working directory. This file is automatically loaded at the start of every conversation, so you don’t have to repeat context:

# Huwise Portal Configuration

## API Access
-Portal URL: <https://myportal.huwise.com>
-Authentication: Use header `Authorization: Apikey YOUR_API_KEY`
-Explore API base: <https://myportal.huwise.com/api/explore/v2.1>
-Automation API base: <https://myportal.huwise.com/api/automation/v1.0>
-Portal API base: <https://myportal.huwise.com/api/portal/v1.0>

## Key datasets
-city-budget-2025: City budget data (1,240 records)
-air-quality-sensors: Real-time air quality (58,302 records)
-public-events: Upcoming city events (412 records)

## Common tasks
-Monthly data quality check: verify all datasets updated within 30 days
-Weekly export: export 'city-budget-2025' as CSV every Monday

This alone can cut your token usage by 30-50% on repeated tasks because you no longer spend tokens re-explaining your setup.

2. Use Environment Variables for Your API Key

Instead of pasting your API key every time, set it as an environment variable. Add this to your shell profile (~/.zshrc or ~/.bashrc):

export HUWISE_API_KEY="your-api-key-here"

Then reference it in your CLAUDE.md:

-Authentication: Use header `Authorization: Apikey $HUWISE_API_KEY`

This is both more secure (the key isn’t in your chat history) and saves tokens.

3. Be Specific in Your Requests

Vague requests waste tokens because Claude Code has to explore before it can act.

4. Use the select Parameter Mentally

When you only need specific fields, say so. This reduces the amount of data Claude Code needs to process:

From the ‘census-data’ dataset, show me only the ‘city_name’ and ‘population’ fields, sorted by population descending, limit 20

This is far more efficient than asking for “all the census data” and then filtering.

5. Batch Related Tasks

Instead of asking five separate questions, combine them:

For the datasets ‘city-budget-2025’, ‘air-quality-sensors’, and ‘public-events’:

1. Show the record count
2. Show the last update date
3. List all field names

One conversation turn instead of five means significantly fewer tokens.

6. Use /compact When Conversations Get Long

If your conversation has been going on for a while and you’re starting a new topic, type /compact in Claude Code. This summarizes the conversation history and frees up context, reducing token usage for subsequent messages.

Tips for a Smooth Experience

• Start small. Try a simple query on the Explore API before moving to creating or modifying datasets with the Automation API.
• Double-check destructive actions. Claude Code will ask for confirmation before deleting or modifying data, but always review what it’s about to do.
• Use natural language. You don’t need to know endpoint paths. Say “delete the test dataset I created yesterday” and Claude Code will figure out the right API call.
• Ask for explanations. If you want to understand what happened, just ask: “What API calls did you just make?” Claude Code will walk you through it.
• Save useful workflows. When you find a task you repeat often, add it to your CLAUDE.md file so future sessions can pick it up instantly.

Quick Reference: Which API Does What?

You don’t need to memorize this — Claude Code picks the right API automatically. But it helps to know the boundaries.

Go further

• Huwise MCP server is here to make most of that even easier. Configure it on Claude

Getting Started Checklist

• [ ] Install Claude Code (npm install -g @anthropic-ai/claude-code)
• [ ] Get your Huwise API key from the back office
• [ ] Create a CLAUDE.md file with your portal configuration
• [ ] Set your API key as an environment variable
• [ ] Try your first query: “List all datasets on my portal”
• [ ] Join the Huwise Community to share tips and ask questions

Please share on the community what you’ve built with Claude!