Claude is a model that analyzes and understands images you upload. It can read and explain photos, charts, screenshots, and documents, or extract text from them — but officially it cannot create images itself. This guide covers how to upload images for analysis and answers the common question of whether Claude can "generate images," based on the official docs. (As of June 2026 · Official: docs.claude.com/vision)
Claude is a model that "analyzes" images
The most important fact first: Anthropic's official docs describe Claude as an "image understanding model." It can interpret and analyze images, but it cannot generate, edit, or composite them. If you came searching for "Claude image generation," know up front that direct generation isn't supported (workarounds are below).
How to upload
In claude.ai, upload an image like a file or drag and drop it into the chat. Supported formats are JPEG, PNG, GIF, and WebP. Per the official docs, you can include up to 20 images at once in claude.ai and up to 100 per API request. The maximum image resolution is 2576 pixels on the long edge (up from 1568 px on prior models). Note that uploaded images are not used to train the models.
What it can do
With vision, Claude can:
- Describe the contents of photos and images
- Read charts, graphs, and tables and interpret the data
- Understand screenshots and scanned documents
- Extract text from handwriting or text within images
- Compare and contrast several images at once
For example, upload a dashboard screenshot and ask for "the three most important trends," or pull text out of a scanned PDF. For long image-based material, it pairs well with document summarizing.
Can it "generate images"?
Not directly. But there are workarounds when you need a visual output.
- SVG, charts, diagrams — build code-based graphics with Artifacts.
- Designs and mockups — canvas-based work lives in Claude Design.
- Connect a generator — link an external image generator via MCP; Claude can write the detailed prompt for that tool.
Limits and cautions
Limitations the official docs note:
- Can't detect AI-generated images — Claude doesn't know if an image is AI-generated and may be wrong if asked. Don't rely on it to spot fakes.
- Medical imaging — it can analyze general medical images but isn't designed to interpret complex diagnostic scans like CT or MRI, and isn't a substitute for professional diagnosis.
- Precision/sensitive analysis — anything needing perfect precision or involving sensitive images requires human oversight.
- Inappropriate images — it won't process explicit content that violates the Acceptable Use Policy.
Tips
- Use clear, high-resolution images; very small or heavily compressed ones reduce accuracy.
- Focus a prompt on the key image rather than dumping many at once.
- Placing the image first, then your question, tends to give better results.
Related
- Summarizing long documents
- Using Artifacts — make SVG/visuals
- Claude Design — the canvas design tool
Disclaimer: This article is based on Anthropic's official Vision documentation (docs.claude.com). The number of images per turn, max resolution, and supported formats may change by model, plan, or version. Check the official docs for the latest. This site is not affiliated with Anthropic.