Enter the name of the Assistant you want to find. The name you enter must match the Assistant's name exactly. For now, we are only able to search your 500 most recently created Assistants.

Enter the filename you want to find, including the file's extension. Example: `my_file_1.csv`. This step will use an exact match to search for that filename.

Optionally filter the results to only search for files with a specific purpose.

Sets the assistant's personality or provides static context.

Refer to the [supported file formats](https://platform.openai.com/docs/assistants/tools/supported-files) to see which tools are compatible with various file formats.

[Supported file formats](https://platform.openai.com/docs/assistants/tools/supported-files). These files will be uploaded to OpenAI and associated with the Assistant.

Use this field to include File IDs of files that have already been uploaded to OpenAI. These files will be associated with the Assistant.

The vector store is a collections of files that can be used to provide context to the assistant. You can use up to 10k files, so if you want your assistant to take advantage of a large number of files, go a head and create a new Vector Store before using this field.

What will this file be used for? Defaults to assistants.

Enter a message or question about your image/s.

Please include the URL of the image. You can add one or more images. GPT-4 access is required. File types supported are PNG (.png), JPEG (.jpeg and .jpg), WEBP (.webp), and non-animated GIF (.gif). Maximum file size is 20 MB. Please enter either a comma-separated list and/or list each URL on a separate line.

The maximum number of tokens to generate in the output.

Enter the text you wish to have analyzed here.

The following fields are for advanced users and should be used with caution as they may affect performance. In most cases, the default options are sufficient. If you'd like to explore these options further, you can [learn more here](https://help.zapier.com/hc/en-us/articles/22497191078797).

The maximum number of tokens to generate in the output. We will do our best to trim old messages to provide this many tokens. Different models may have different total maximums, which include all instructions & messages, and generated output (IE: 8192 for [gpt-4](https://platform.openai.com/docs/models/gpt-4) or 4096 for [gpt-3.5-turbo](https://platform.openai.com/docs/models/gpt-3-5)).

GPT-4 not available? **[Learn how](https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4)** to enable it and **[set up a paid OpenAI API account](https://platform.openai.com/account/billing/overview)** (if necessary).

A number between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. Recommend altering only **Temperature** or **Top P** (and not both).

Nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. Recommend altering only **Temperature** or **Top P** (and not both).

Data sent to OpenAI through this Zap is via API. Under OpenAI's [API data usage policy](https://openai.com/policies/api-data-usage-policies), OpenAI will not use API-submitted data to train or improve their models unless you explicitly decide to share your data with them for that purpose (such as by opting in). For more information, please review OpenAI's article about [when/how data may be used to improve model performance](https://help.openai.com/en/articles/5722486-how-your-data-is-used-to-improve-model-performance).

When testing the action, it will run for 50 seconds. If the entire response is not available, the `was_cancelled` output field will be `true`.

The most recent message the assistant (ChatGPT) responds to.

If provided, this unique value will allow the assistant to continue a conversation from previous messages. Use something that won't change over the course of a conversation (like a user's email address). If blank, the assistant will not remember this conversation going forward.

Please include the URL of the image. If an image is sent, you must use the model `gpt-4-vision-preview`. GPT-4 access is required. File types supported are PNG (.png), JPEG (.jpeg and .jpg), WEBP (.webp), and non-animated GIF (.gif). Maximum file size is 20 MB.

The name of the user sending a message to the assistant. *Feel free to experiment with this!*

The name of the assistant receiving the message. *Feel free to experiment with this!*

Instructions to the assistant. *Feel free to experiment with this!*

This action will classify the text provided and determine if it is potentially harmful. The classification will be based on the [OpenAI moderation guide](https://platform.openai.com/docs/guides/moderation/overview).

Learn more about [moderation models compatibility](https://platform.openai.com/docs/models/model-endpoint-compatibility)

Enter the text you wish to have classified here.

Please enter either a comma-separated list and/or list each category on a separate line.

The message you want ChatGPT to respond to. Not sure what to write? [Check out this article.](https://zapier.com/blog/gpt-prompt/)

If you leave this empty, we will create a new Assistant each time your Zap runs using the fields below.

Include background information, writing style, or other helpful details to give ChatGPT the right context for the conversation.

Defaults to gpt-3.5-turbo. If you have selected an existing Assistant, this will override that Assistant's model. If you have not selected an Assistant, we will use this model when we create a new Assistant.

Refer to the [supported file formats](https://platform.openai.com/docs/assistants/tools/supported-files) to see which tools are compatible with various file formats. If you have selected an existing Assistant, this field will override that Assistant's tools. If you have not selected an Assistant, we will use this field when we create a new Assistant.

If provided, this unique value will allow the assistant to continue a conversation from previous messages. Use something that won't change over the course of a conversation (like a customer support ticket ID.) If blank, the assistant will not remember this conversation going forward.

[Supported file formats](https://platform.openai.com/docs/assistants/tools/supported-files). These files will be uploaded to OpenAI and associated with the Message. They will not be associated with the Assistant.

Use this field to include the File IDs of files that have already been uploaded to OpenAI. Any files included here will be associated with the Message. They will not be associated with the Assistant.

These instructions override the Assistant Instructions and can be specific to each message. Helpful if you want specific instructions per message.

[Supported file formats](https://platform.openai.com/docs/assistants/tools/supported-files). These files will be [uploaded to OpenAI](https://platform.openai.com/docs/api-reference/files/create) and associated with the Vector Store. The `purpose` property of these files will be set to `assistants`.

Use this field to include comma-separated File IDs for files that have already been uploaded to OpenAI. These files will be associated with the Vector Store. Example: `file-1234, file-5678`

Provide the unstructured text (any text content, it can be very messy!) from which you want to extract structured data. This is where you will want to provide mapped values from previous steps!

Select the model you'd like to use. Note: only the *-0613 models (and beyond) have support for [function calling](https://platform.openai.com/docs/guides/gpt/function-calling), which is required for this action.

A number between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. For data extraction, we recommend using a very low temperature value like 0.0.

Provide a brief description of what sort of data you want extracted from the unstructured text. Provide general tips or tricks to help guide the AI.

Provide the names of the values you want to extract from the unstructured text. Names should be unique and short. You may use "." to indicate nested values. After you provide the names, you will be able to provide additional configuration options for each value below.

A text description of the desired image(s). The maximum length is 1000 characters for dall-e-2 and 4000 characters for dall-e-3.

The number of images to generate. Must be between 1 and 10. For dall-e-3, only n=1 is supported.

A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse. [Learn more](https://platform.openai.com/docs/guides/safety-best-practices/end-user-ids)

Enter the text you wish to have summarized here. We recommend a maximum of 7000 characters.

Select the model you'd like to use. We recommend `gpt-3.5-turbo` for this action. GPT-4 not available? **[Learn how](https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4)** to enable it and **[set up a paid OpenAI API account](https://platform.openai.com/account/billing/overview)** (if necessary).

The maximum number of tokens to generate in the output. Different models may have different total maximums, which include all instructions, messages, and generated output (IE: 8192 for [gpt-4](https://platform.openai.com/docs/models/gpt-4) or 4096 for [gpt-3.5-turbo](https://platform.openai.com/docs/models/gpt-3-5)).

The audio file object (not file name) to translate, in one of these formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.

An optional text to guide the model's style or continue a previous audio segment. The prompt should be in English.

Enter the name of the person to whom you want to send an email.

Enter the email address of the person to whom you want to send an email.

Provide a detailed prompt describing the content of the email you wish to send.

Select the tone of the content of the email.

API docs for ChatGPT are located here https://platform.openai.com/docs/api-reference/chat

Requests can only be made to domains associated by this application. View [documentation](https://help.zapier.com/hc/en-us/articles/12899607716493) for more details.

These params will be URL-encoded and appended to the URL when making the request.

Authentication headers are included automatically

Zapier will apply these optional headers and values to the request in addition to any headers created as part of the authentication mechanism.

Zapier will pass this along raw. If you would like to provide complex JSON, try a "Code by Zapier" step to build the structure and pass it along here. For example, `{"foo": "bar"}` is perfectly valid.

ChatGPT is an AI tool built by OpenAI that enables conversational automation.

A more powerful OpenAI integration is available here: https://zapier.com/apps/openai/integrations

Zapier Automation Platform

PRODUCTS

CAPABILITIES

By use case

By app

By team

By company size

How Zapier Zaps

How Zapier Uses AI

By team

Learn more

Get help

Zapier quick-start guide

Developer resources

ChatGPT Integrations

ChatGPT Help & Support

OpenAI and ChatGPT error: Insufficient_quota

Advanced fields for OpenAI and ChatGPT

Use OpenAI's Assistants API with your Zaps

How to get started with ChatGPT on Zapier

ChatGPT Tutorials

How to build an analysis assistant with ChatGPT

How to automate your event promotion with Zapier and ChatGPT

Automate document outlines with ChatGPT and Zapier

About ChatGPT

Related categories