ChatGPT

WorkOnGPT.com provides a comprehensive portal to a variety of smart office services, including document conversion, document translation, document Q&A, and more. One of the most prominent features is the ChatGPT interaction. Here is a detailed guide on how to use the ChatGPT feature:

Talk with ChatGPT

• To start a conversation: Simply type your question into the text box provided and click "Send" or press "Enter". You can also directly select the AI character you want to talk to and ask a question later.

• Continuous Interaction: ChatGPT remembers the context of a conversation up to its model limit to ensure contextual coherence.

• Historical conversations: The platform automatically saves your current conversations. You can view past conversations, return to previous chats, and continue from where you left off. You can give each conversation a customized title for reference.

• Deleting Conversations: For privacy reasons, we offer the option to "Delete Conversations". Once a history conversation is deleted, it will be removed from both the frontend and the server. You can delete previous conversations one by one or in batches.

Talk with Microsoft's Copilot

This site also provides the AI networking search function for Microsoft's Copilot. You can select a Copilot model, enter your question, and the AI will collect relevant information from the web and return reference links to help you get accurate and timely answers. Please note that talking to Microsoft Copilot does not support mobile response.

Modeling Capability

We will update the types and capabilities of the models according to the OpenAI official documentation to ensure that your experience is consistent with the capabilities of the official models.The capabilities of the various ChatGPT models on WorkOnGPT.com are listed below:

AI Model	Memory (Number of Tokens)	Max Reply(Number of Tokens)
gpt-4o-mini	4000	3500
gpt-4o	5000	2000
gpt-4-turbo	5000	2000
ChatGLM4-9B	5000	2000
Copilot	1 Message	N/A

The Copilot model allows you to perform AI networked searches, and since Microsoft has not released an official API, this site uses a wrapped API port as a way to provide you with this functionality, as well as providing 1 historical conversation as a memory space for Copilot.

In addition, we provide the ChatGLM4-9B model, ChatGLM4-9B is an open-source, 26-language-supporting AI model based on the General Language Model (GLM) architecture with 6.2 billion parameters. This model is jointly developed by Wisdom Spectrum AI and a team from Tsinghua University. This site has been granted a commercial license for ChatGLM4-9B, and the ChatGLM4-9B model is deployed through localization for free use by users of this site.

Advanced Features

1. Export Dialogs: You can export your dialogues at any time to a docx document that will contain the website address and a short introduction to the site.

2. Extended Models: For more tedious queries or broader contexts, you can choose between different GPT models, including the GPT model and the ChatGLM model. You are free to choose which model to use in the same round of dialogue, and the AI retains the memory of the dialog context after changing models.

3. Role Setting: This site has built-in AI characters, such as AI doctor, AI chef, AI lover, AI Tarot Master and other characters. In addition, you can customize the system prompt according to your own usage habits, so as to better meet your work needs. To set up a customized role, please tell the AI in the first person what role it is playing, what tasks it needs to accomplish, and what kind of format or style the reply is in. After selecting the role, you cannot switch to the Copilot model, but you can switch to other ChatGPT and ChatGLM models.

Safety & Privacy

• We take your privacy very seriously. Your conversation data is stored on our servers and encrypted with SHA-256.

• By encrypting your message data, even server backends cannot access user conversations.

Billing

• Using the GPT model, each conversation is billed based on the token used.

• You can choose whether or not to top up your membership on demand, and when you have a conversation with ChatGPT, the corresponding fee will be deducted from your account balance.

• For the Microsoft copilot model, the cost of asking a question is $0.03 per conversation.

• ChatGPT feature with ChatGLM4-9B model has been completely free, come and try it.

Please note that every 1000 tokens corresponds to roughly 750 characters. According to our tests, the maximum number of tokens returned by each model in a single reply is generally around 1,000 tokens for the official API interface of ChatGPT, which is lower than our required limit of tokens for a single reply. You don't need to worry about the problem of model replies being truncated. For the billing rate of each model, please see the website policy section of this article (click to jump to the pricing policy).

WorkOnGPT.com does not only provide AI large model interaction services. We have designed a series of derivative functions for your professional office needs, such as document translation, AI article writing, voice recognition, AI search and other functions, to help you complete all kinds of work more efficiently.