Working with Files in ChatGPT

ChatGPT allows users to upload and work with files directly within conversations to analyze spreadsheets, edit documents, summarize PDFs, and process images. This capability enables users to perform complex data analysis and content refinement without leaving the chat interface.

File Upload and Processing Workflow

ChatGPT supports a variety of file formats, including CSV, XLSX, PDF, DOCX, JPEG, PNG, and TXT. The workflow for working with files involves starting a chat and selecting "Add photos or files" from the tools menu to upload the necessary documents or images.

Once a file is uploaded, users can perform several types of tasks:

• Summarization and Analysis: Users can request summaries of reports, identifying specific risks or open questions.
• Data Visualization: ChatGPT can visualize sales data by region and highlight month-over-month changes.
• Content Editing: The system can rewrite documents for clarity and conciseness while maintaining the original tone.
• Data Extraction: Key information, such as dates and owners, can be extracted from PDFs into structured tables.

Users can further refine their results by requesting specific views, such as charts or tables, and can download the outputs generated by ChatGPT, including updated spreadsheets or PDFs.

Third-Party App Integrations

Certain versions of ChatGPT support apps, which allow the model to access third-party tools to bring external context into the conversation. These apps are managed via the Settings → Apps menu, where users complete an authentication and permissions flow. Once connected, apps can be accessed through the Tools menu, or by using the @ or / commands during a chat.

For Enterprise and Business workspace users, the following conditions apply:

• Administrative Control: Organization admins control which apps are available to users.
• Data Privacy: Data accessed through apps in ChatGPT Enterprise or ChatGPT Business workspaces is not used to train OpenAI models by default.