AnythingLLM has no hard file size limit, but large PDFs cause CPU exhaustion and 502 errors. Hosted Starter tier recommends under 10k words per file; Professional tier under 50k. Compress yours here before uploading.
⚡ AnythingLLM File Size Limits (Quick Fix)
•Docker deployments of AnythingLLM behind an NGINX reverse proxy default to a 1MB upload limit via client_max_body_size, which will silently drop larger files.
👉 Fix: Compress your file below the required limit using the tool above.
📄
Drop your PDF here
or click to upload — PDF only
Target size (optional)
or leave blank to compress as much as possible
How to use this tool
Select or drag and drop your file into the tool above.
Adjust the settings or target size as needed for your specific requirement.
Wait a moment while your file is processed directly in your browser.
Download the final file safely to your device.
About this tool
Local and Docker installs have no enforced limit but can crash on very large PDFs due to RAM or CPU exhaustion. Docker deployments behind NGINX default to a 1MB upload cap — this must be raised via client_max_body_size. For best results, split large PDFs into chapters and consider converting to Markdown for faster embedding.
→AnythingLLM hosted Starter tier recommends a maximum of 10,000 words per PDF file to avoid CPU exhaustion and 502 errors.GitHub ↗
→AnythingLLM hosted Professional tier recommends a maximum of 50,000 words per PDF file.GitHub ↗
→Local and Docker installs of AnythingLLM have no enforced file size limit, but very large PDFs can exhaust CPU or RAM and cause the process to crash.GitHub ↗
→Docker deployments of AnythingLLM behind an NGINX reverse proxy default to a 1MB upload limit via client_max_body_size, which will silently drop larger files.GitHub ↗
→Converting large PDFs to Markdown before uploading to AnythingLLM improves embedding performance, as PDFs are less efficient to parse than plain text formats.GitHub ↗
❓ Frequently Asked Questions
Does AnythingLLM have a PDF file size limit?+
No hard limit is enforced, but practical limits exist. The hosted Starter tier recommends under 10k words per file; Professional tier under 50k. Local/Docker installs can crash on very large PDFs due to CPU or RAM exhaustion.Source: GitHub ↗
What causes 502 errors when uploading PDFs to AnythingLLM?+
502 errors on the hosted version are caused by CPU exhaustion from processing very large PDFs. Keep files under 10k words on the Starter tier and under 50k words on the Professional tier.Source: GitHub ↗
Why is AnythingLLM dropping my PDF upload silently?+
If you're running AnythingLLM in Docker behind NGINX, the default client_max_body_size is 1MB. Files larger than 1MB are silently dropped. Increase this limit in your NGINX config.Source: GitHub ↗
What is the best way to handle large PDFs in AnythingLLM?+
Split the PDF into smaller chapters, compress it to reduce size, or convert it to Markdown. Markdown is faster to embed and produces better AI responses than large scanned PDFs.Source: GitHub ↗
Does compressing a PDF affect AnythingLLM's ability to read it?+
No. AnythingLLM reads the text layer of the PDF. Compressing image quality reduces file size without affecting the text content used for embedding and chat.Source: GitHub ↗