Fix Claude Not Reading Uploaded PDFs: Step-by-Step

When Claude fails to read or analyze an uploaded PDF, it usually displays a generic error message, ignores the document's contents, or hallucinates a response based only on your prompt text. Because Claude relies on internal parsing pipelines to convert documents into plain text before processing, any formatting, security, or size anomalies in the PDF can break this chain.

Use these troubleshooting steps to identify why Claude cannot read your PDF and fix the formatting so the model can process it.

1. Verify if the PDF is scanned (Image-only) Claude can struggle to read PDFs that are actually collections of scanned images rather than selectable text. 1. Open the PDF file on your computer using any standard viewer (like Adobe Acrobat, Preview, or a web browser). 2. Try to click and drag your cursor to highlight a sentence. 3. If you cannot select individual words or characters, the PDF is an image. 4. Run the file through an Optical Character Recognition (OCR) tool (such as Adobe Acrobat's PDF Auto-OCR, or free online converters) to generate a searchable text layer, then re-upload the new file to Claude.

2. Reduce the file size and page count While Claude has a massive context window, the document upload pipeline has strict limits on individual file sizes and processing overhead. 1. Check the file size of your PDF. If it exceeds 10MB, Claude may fail to ingest it. 2. Compress the PDF using a tool like Adobe PDF Compressor to shrink the file size without losing text quality. 3. If the document is exceptionally long (e.g., hundreds of pages), split the PDF into smaller sections of 20 to 50 pages each. Upload only the relevant sections to keep the processing token load low.

3. Remove security permissions and password protection Claude cannot decrypt password-protected files or process PDFs with strict digital rights management (DRM) restrictions. 1. Open your PDF and check if it prompts you for a password. If it does, you must save a decrypted copy. 2. Check the PDF document properties/permissions. If "Content Copying" or "Extraction" is set to "Not Allowed," Claude's parser will be blocked. 3. To fix this, print the document to a new PDF file ("Save as PDF" or "Print to PDF"). This flattens the file and strips security restrictions.

4. Convert the PDF to Markdown or TXT Complex layouts, multi-column academic papers, tables, and nested charts can confuse Claude's PDF ingestion tool, leading to failed reads or garbled context. 1. Open your PDF and export or save it as a plain text (`.txt`) or Markdown (`.md`) file. 2. Review the exported text file to ensure the structure makes sense and formatting characters have not corrupted the words. 3. Upload this `.txt` or `.md` file directly to Claude. These formats skip the PDF parser entirely and are read cleanly by the model.

5. Check your browser and session state Sometimes, local browser memory leaks or network drops interrupt the file upload process before it completes. 1. Clear your browser cache and cookies, or open an Incognito/Private window. 2. Re-log into your Anthropic account. 3. Ensure your internet connection is stable, then upload the file again. Wait for the upload progress bar to fully complete before typing your prompt and clicking submit.

When to escalate If you have converted the file to plain text, confirmed it has no passwords, kept it under 5MB, and Claude still throws an "error reading file" message, check the Anthropic Status page. System-wide degradation of the document processing pipeline can occur during high-traffic periods. If the status page reports all systems are functional, contact Anthropic Support via the in-app help widget, providing the file size, format, and the specific error message you received.