IndexOpenAI chatbot
OpenAI chatbot
An OpenAI chatbot is your interface on top of the API: a system instruction, the message list, and a model id. ChatGPT the website is the hosted version of that idea, not a widget you embed.
Updated 2026-09-27
The community walkthrough people still open is Creating a chatbot with the OpenAI API. The current reference is the API docs. Create a key on the API account.
A small version that holds up
- Write the instruction as a job: what it may answer, and what it must refuse
- Send the conversation so far, not only the latest line, or it will forget
- Store the key in the server environment. The page the user sees calls your server
- Log token usage from the response before you scale. The token counter is only a preview
- Add tools or handoffs in Agent Builder when a single reply cannot do the task
Questions
Can I embed ChatGPT on my site?
You call the API and render your own UI. You do not iframe chatgpt.com. Brand rules for the logo are on openai.com/brand.
Which model do I pick?
A current small model for cheap drafts, a larger one when the answer has to be right. Read the model list the day you ship. GPT-5.4 left the ChatGPT app but API ids are their own list.
Related notes
- OpenAI API and API keys
Create an OpenAI API key at platform.openai.com. Usage is billed apart from ChatGPT Plus. new-api is a separate open-source gateway, not the official API.
- OpenAI Agent Builder
Agent Builder is the visual canvas in the OpenAI API docs for composing agents, tools, and handoffs. It is not the ChatGPT Agent button.
- Token counter
Estimate an OpenAI token count at about four English characters per token, then confirm on platform.openai.com/tokenizer before you budget a prompt.
- ChatGPT for customer service
How support teams use ChatGPT to draft replies, and where an agent or a workspace policy fits. This is not the OpenAI help desk.