ChatGPT's file and image uploads let you turn messy documents, spreadsheets, and photos into actionable insights. You can ask it to summarize a research paper, extract data from a spreadsheet, compare two presentations, or describe what’s in a photo.
This guide walks you through the features, limits, and best practices to make the most of multimodal analysis.
File uploads allow you to upload documents such as PDFs, Word files, PowerPoint presentations, spreadsheets (CSV/XLSX), and plain text. Once uploaded, ChatGPT can summarize the content, extract specific information, compare multiple documents, or help rewrite sections in different styles.
Image uploads enable you to add photos, diagrams, screenshots, or charts and ask questions about objects in the image or analyze visual content.
Most common file types are supported, including PDFs, DOCX, PPTX, text files, CSV/XLSX spreadsheets, and images (PNG, JPEG/JPG, GIF). Key limits to know:
To upload a document, open a ChatGPT conversation and click the + icon in the prompt area (or the paperclip if available). Select your file or drag it into the chat window. Wait for the file to finish uploading before submitting your message. ChatGPT will ingest the content and respond with insights or actions you request, such as:
Try it now: Upload a PDF (under 512MB) and ask ChatGPT to summarize the key points in three bullet points.
Try it now: Upload a CSV or spreadsheet and request a visualization of trends or a summary of totals.
Image uploads work similarly. Tap the + icon and choose Add photos & files, drag an image into the prompt area, or paste a copied image.
Then ask ChatGPT questions about objects, text, or patterns in the image. You can also annotate the image beforehand (e.g., circle a region) to guide ChatGPT.
Try it now: Upload a photo and ask ChatGPT to identify the objects it sees.
Try it now: Take a screenshot of a chart and request an explanation of trends, labels, and axes.
Chats and files remain in your account until you delete them. When you delete a chat or your account, associated files are removed from OpenAI’s systems within 30 days.
Files uploaded as knowledge for custom GPTs persist until you delete the GPT. OpenAI may use content (including files and images) from non-enterprise customers to improve model performance; you can opt out by adjusting your data settings.
Try it now: Delete a chat that contains an uploaded file and confirm that the file is removed from the context.
Try it now: Review your data settings and opt out of training if desired.
Combine multiple modalities for a richer analysis:
Experiment with different file types and prompts. By iterating, you’ll learn how to guide ChatGPT effectively.
File and image uploads expand ChatGPT’s capabilities beyond simple text conversations. When used thoughtfully, these tools can help you synthesize complex information, extract actionable insights, and explore visuals interactively. Keep file sizes in mind, protect sensitive information, and experiment with prompts to get the most value out of multimodal analysis.
1 Tasks enabled by file uploads—synthesis, transformation, and extraction—are described in the File Uploads FAQ.
2 Limits on file numbers, sizes, and quotas come from the File Uploads FAQ.
3 Image upload instructions and supported formats are based on the Image Inputs FAQ.
4 Limitations and cautionary notes for image analysis appear in the Image Inputs FAQ.
5 General description of ChatGPT’s ability to analyze uploaded documents and images is provided in the Capabilities Overview.