[BUG] PDFs are limited to 256KB and No Vision/OCR is performed with AWS Bedrock

Resolved 💬 3 comments Opened Dec 21, 2025 by r3core Closed Feb 14, 2026

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

When using Claude Code or Claude Agent SDK with AWS Bedrock and Attempting to read PDFs with image data.

  • Read tool fails for any PDF larger than 256KB
  • Smaller PDFs get read, but no multimodal vision analysis is performed and Claude only sees raw PDF data with objects and streams.

What Should Happen?

The Bedrock flow should either have feature parity with the Anthropic API flow and allow us to perform actual analysis on PDF documents. OR the documentation should clearly state the Claude Code/Claude Agent SDK with Bedrock API has this limitation regarding PDFs and what are the alternatives.

Error Messages/Logs

Steps to Reproduce

  1. Start Claude Code with AWS Bedrock API credentials
  2. Run it on a folder with PDFs with image data
  3. Ask Claude to analyse

Claude Model

Sonnet (default)

Is this a regression?

No, this never worked

Last Working Version

_No response_

Claude Code Version

2.0.64

Platform

AWS Bedrock

Operating System

macOS

Terminal/Shell

Other

Additional Information

<img width="877" height="612" alt="Image" src="https://github.com/user-attachments/assets/f59c30fa-d446-4d87-a4e9-79b46b49810b" />

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗