Custom GPTs: Create a Custom ChatGPT with Your Data

Learn to go beyond the GPT Builder by looking under the hood, improving instructions, exploring conversation starters, and sharing advanced GPTs, while understanding non-determinism and model updates.

Fine-tune your custom GPT by crafting specific, structured instructions in the GPT Builder, using instant updates and preview mode to add preferred ingredients and prep time.

Save changes by clicking update and confirm, as unpublished edits won't auto save. Use backups and the revert option to restore the last saved version when edits diverge.

Learn how conversation starters, predefined prompts configured in the GPT builder, help you quickly start interactions by editing, adding, or deleting them so only the first four appear.

Learn to update a GPT's profile picture via the GPT builder or the configure option, generate a veggies-only image, and optionally upload or use DALL-E, noting possible delays.

Learn to create and configure a custom GPT from scratch, build a virtual support agent for the AI Webcam product, and craft precise instructions with conversation starters and troubleshooting tips.

Understand prompt context across create, preview, and live modes in the GPT builder, test prompts, provide feedback, and update instructions to fine-tune and troubleshoot your custom GPT.

Configure your GPT to keep internal configuration and secret data confidential by instructing it not to reveal secrets, even in preview, while noting platform changes may affect behavior.

Share your custom GPT via a link in the GPT store; users must sign in to OpenAI and have a ChatGPT Plus subscription, and embedding on websites isn’t supported.

Learn to manage your builder profile, hide or set your creator name from billing details, and verify a domain (dns changes) to drive traffic to your website via public GPTs.

Learn to delete a GPT by using Edit GPT or the Explore view, then click the three dots and choose delete. This action cannot be undone, so save configurations elsewhere.

Explore the core capabilities you can enable for GPTs, including web browsing, image generation with DALL-E, and code execution, and learn when to keep or adjust them.

Learn how to enable and use web browsing in a custom GPT to access up-to-date information, overcome knowledge cutoff limits, and ensure responses pull current data.

Enable the DALL-E image generation capability in your GPT and use engineered prompts to generate recipe visuals for the Nutri Chef GPT, with options to disable if it fails.

Explore the code interpreter that uses Python to write and execute code, analyze data, plot a sine wave, and transform json to yaml with file handling.

Explore how to provide your own files to a custom GPT, overcoming ChatGPT cut off date and enabling you to chat with your data through analysis and visualization.

Upload PDFs in the knowledge section to power a custom GPT, then reference the provided document in answers. Max ten files, total size 100 MB.

Explore types of queries for document-based knowledge in custom GPTs: searching my knowledge and analyzing, with the code interpreter processing uploaded files and extracting insights.

Start small with a single text-based document type (PDF, Word, spreadsheets) to avoid processing errors and delays. Double-check results and gradually add more; large files and images often cause issues.

Analyze a spreadsheet by uploading a csv, enabling code interpreter to process sales data, and identify the best selling product and the best and worst months.

Explore visualizing spreadsheet data with the code interpreter by generating a Python-based bar chart of monthly sales, with Dall-E disabled and data extraction revealing January and December trends.

learn how a code interpreter reads user documents and generates a new translated German document from English, with an iterative, self-correcting, download-ready translation workflow.

Protect privacy by uploading a file and disabling 'use conversation data in your GPT to improve our models' in additional settings, preventing data from being used to train OpenAI models.

Explore common questions about uploading files for custom GPTs, including size limits and the ten-file save rule. Learn how to optimize formats, from txt to json, for faster processing.

Learn how to extend custom GPTs with actions that perform API calls to send emails, fetch data, and automate tasks, turning GPTs into practical agents.

An api is an interface for computers to communicate with a server, unlike a human-facing gui; a custom gpt can call one or more apis to perform actions.

Learn to configure a custom GPT to call an API via GPT actions, import an OpenAPI schema, test actions, and build a vehicle maintenance quote workflow.

Explore the open API specification as a standard language for describing API structures, including schemas, an operation ID, and endpoints, and how GPT uses them to determine actions.

Learn how Zapier acts as an automation tool to integrate GPTs with over 6000 apps through a friendly UI, enabling Zapier AI in your GPT.

Learn to connect a custom gpt with Zapier ai actions to add events to Google Calendar by importing an api spec and configuring actions.