<aside>

On this page:

</aside>

<aside>

Expanse Docs

🧠 Product Docs

🤖 Prompting Tips

LLM Directory

Prompting Tips & Tricks

Prompt Library

💻 Use Cases

For Work

For Fun

📚 Resources

Prompt Engineering Resources

🍄 About Expanse

Meet the team

</aside>

The Difference Between Using ChatGPT’s Platform vs. API

If you've been using ChatGPT through their website or mobile app, you might have come across some strange behaviors such as a change in tone-of-voice, mysteriously remembering or referencing chats that you deleted, or limiting your usage due to message, image or content restriction limits.

While these pain points are prominent across ChatGPT and other AI models via their official providers, they are easily avoidable when you access these models through an app that hosts their API. These apps let you access the same AI but with more control, better privacy, and often at a lower cost, making them the better alternative.

<aside> 💡

TLDR

  1. System prompts: ChatGPT controls the invisible instructions that shape responses; API apps let you customize them
  2. Privacy: Your ChatGPT conversations are stored on their servers; API apps can keep data on your device
  3. Cost: $20/month flat fee vs. pay-per-use pricing that scales with your actual usage
  4. Performance: Shared resources and limits vs. dedicated performance and fewer restrictions

</aside>


🌫️ The Hidden System Prompt

Have you ever noticed your AI suddenly change its tone-of-voice? Or tell you it can do one thing one day, but refuse to do it the next? Well, this all comes down to the system prompt- a hidden filter that AI providers run any of your messages through before answering.

What is a system prompt and why does it matter?

A system prompt acts as an invisible set of instructions that guides how the AI responds to you. Think of it as the "personality settings" for the AI. OpenAI can change this system prompt at any time without telling you, which can:

<aside> 💡

These hidden instructions can override your specific requests, which is why ChatGPT sometimes refuses tasks that should be perfectly reasonable.

</aside>

But if you use ChatGPT through API…

When you use ChatGPT through the API with an app like Expanse, there is no system prompt, so the responses remain direct and unfiltered.

Instead, you are able to generate your own system prompts through features like Roles where you can specify:

<aside> 💡

You can create multiple Roles to make it easier to switch between modes depending on your task at hand.

</aside>


🔐 Privacy and Security

The way you access ChatGPT significantly impacts who can see (and control) your data.

Using ChatGPT Platform vs API Apps like Expanse

When using the official ChatGPT interface When you use an API-based app like Expanse
Your data lives in the cloud Your data stays with you
All your conversations are linked directly to your account All your conversations are stored locally on your device
Your chat history is stored on OpenAI's servers Your data isn't automatically linked to your personal identity
Your data is automatically used for model training (unless you manually opt out) Conversations aren't used for model training
Your conversations may be accessible to OpenAI staff You maintain control over your information

<aside> 💡

Why keeping your data local matters: If you're discussing sensitive business information, personal projects, or confidential data, keeping your conversations local gives you an extra layer of privacy protection.

</aside>


💰 Subscription Costs

How you access ChatGPT doesn't just affect your experience- it affects your wallet too.

ChatGPT Plus: One-size-fits-all pricing

With the official ChatGPT Plus subscription:

Using ChatGPT through Expanse: Transparent, usage-based pricing

With API access through apps like Expanse: