Day 2 Exercise: Run the Same Request on an Aggregator Platform

This is the Day 2 companion exercise. Complete Day 1 first.

Yesterday you ran your first API call through DeepSeek’s official API. Today we do one thing: switch to a different platform, change two parameters in the same code, and run it again.

You’ll see that learning one platform’s API means you’ve learned them all—as long as they’re compatible with the OpenAI interface.

What Is an Aggregator Platform

An aggregator platform is a middle layer. You register one account, top up once, and get access to dozens of AI models (OpenAI, Anthropic, Google, etc.) without signing up at each official platform separately.

Under the hood, most aggregators run on NewAPI or OneAPI frameworks, so the workflow is nearly identical across platforms:

Register an account
Top up your balance
Get your API Key
Pick a model from the model list

Pros: One key for many models, no need to register at each official platform separately, supports local payment methods.

Cons: Slightly more expensive than official APIs (platform service fee on top), stability depends on the aggregator.

“Write” Your Code

No need to start from scratch. Copy yesterday’s hello_api.py as hello_aggregator.py and change exactly two things:

import os
from dotenv import load_dotenv
from openai import OpenAI

load_dotenv()

# Change 1: Use the aggregator's key
api_key = os.environ.get("AGGREGATOR_API_KEY")
if not api_key:
    raise ValueError("AGGREGATOR_API_KEY not set. Check your .env file.")

# Change 2: Use the aggregator's API endpoint
client = OpenAI(
    api_key=api_key,
    base_url="https://your-aggregator-url/v1"
)

# Everything else stays the same
response = client.chat.completions.create(
    model="deepseek-v4-flash",       # Model name may differ on aggregator—check their docs
    messages=[
        {"role": "user", "content": "Hello, introduce yourself in one sentence."}
    ]
)

print(response.choices[0].message.content)

Only two changes: api_key and base_url. The logic is identical.

Add the aggregator key to your .env file:

DEEPSEEK_API_KEY=sk-xxx
AGGREGATOR_API_KEY=sk-your-aggregator-key

Run it:

uv run python hello_aggregator.py

AI replied? Same code, two parameters changed, different platform. That’s the benefit of OpenAI-compatible interfaces.

Things to Watch Out For

Model names may differ

The model name on an aggregator might not be deepseek-v4-flash. Some use deepseek/deepseek-v4-flash (with a provider prefix), others use different formats. Check the aggregator’s model list or pricing page for the correct name.

base_url trailing slash

Aggregators usually need a /v1 suffix, e.g. https://example.com/v1. Official APIs (like DeepSeek) usually don’t need it. When in doubt, check the aggregator’s documentation for the correct API endpoint format.

When in doubt, ask AI

In L0→L1 you learned how to chat with AI—now put that to use. Not sure about something? Ask in a chat window: “How do I set the base_url for [platform name] using the Python openai library?” or “What’s the model name for deepseek-v4-flash on this platform?” It’s faster than digging through docs yourself.

Comparison: Aggregator vs Official API

	Official API (e.g. DeepSeek)	Aggregator Platform
Registration	Separate account per platform	One account, many models
Payment	Top up at each platform separately	Top up once, pay per model
Pricing	Official rate	Discounted rate (typically cheaper than official)
Model selection	Only their own models	Dozens of models to switch between
Stability	Depends on the provider	Depends on both aggregator and provider
Best for	Long-term use of a specific model	Trying out multiple models without separate signups

Recommendation: Use an aggregator during the learning phase to try different models, then switch to the official API for your primary model to save on costs.

Troubleshooting

“Model not found”

Model names differ from official APIs—check the aggregator’s documentation

“Invalid API key”

Make sure you’re using the aggregator’s key, not DeepSeek’s

“Insufficient balance”

Top up your aggregator account and retry

“Connection refused”

Check base_url: make sure it has /v1 suffix (aggregators usually need it)
Double-check the domain spelling

What You Did Today

Registered on an aggregator platform and got an API Key
Changed two parameters in yesterday’s code and ran it on the aggregator
Understood the difference between aggregator and official APIs

Next up: Day 3 is all about parameters—turn temperature from 0 to 1 and see how the AI’s responses change.

What Is an Aggregator Platform#

“Write” Your Code#

Things to Watch Out For#

Comparison: Aggregator vs Official API#

Troubleshooting#

What You Did Today#