HARPA.AI
USE CASESGUIDESAI COMMANDSBLOGPRIVACY

Connecting to CloudGPT

CONTENTS
NAVIGATE

Cloud GPT is our premium connection. It has significantly less downtime and isn't restricted by country, ensuring availability even when OpenAI's servers are at capacity.

You are billed based on your usage, measured in Megatokens, which can be purchased regardless of your subscription plan. S plan users receive several Megatokens as part of their package.

# How to use CloudGPT?

  1. If you purchased the S plan, you already have Megatokens. X or Free plan users can buy Megatokens by:

  1. Open HARPA by clicking its icon in the top right corner of the Chrome Browser or by pressing Alt+A (on Windows) and ^+A (on Mac).

  2. Create a new chat and you'll see the default AI model and your current CloudGPT token balance (if CloudGPT connection is selected):

  1. Alternatively: click on the connection menu in the bottom left corner, and you'll see available CloudGPT connection options and your current Megatoken balance at the top.

Usage control:

  • Check your chat request's input and output tokens by hovering over the star icon below the message for details.

  • Usage history could be checked from your space profile on HARPA AI.

This can be useful for Spaces users, where the Megatoken balance is shared, monitored, and paid for by the owner.

# What are Megatokens?

1 Megatoken is 1 million GPT-4o-mini input tokens and is roughly 750.000 words of input text in English. Using languages with non-Latin alphabets significantly increases the number of tokens used.

Output tokens cost x3 input tokens. GPT-4o tokens cost x10 of GPT-4o-mini tokens.

How many queries is in 1 Megatoken?

On average, 1 Megatoken is:

  • Approximately 150 YouTube summaries (based on 15-minute videos) using GPT-4o-mini. GPT-4o or Sonnet models would use 10 times more tokens.
  • Roughly 50 GPT-4o-mini or GPT-4o Vision requests.
  • 3000 basic requests in chat on GPT-4o-mini model.
  • 300 page-aware requests in chat on GPT-4o-mini model.
  • 300 basic requests in chat on GPT-4o model.
  • 30 page-aware requests in chat on GPT-4o model.
NEXT POST
Contact us
HomeUse CasesGuidesPrivacy PolicyTerms of Service