Fix a failed PDF summary.

PDF summary failures fall into three buckets: extraction failed (the PDF has no selectable text — usually an image-only scan or encrypted document), the source is too large (50+ page documents that exceed the AI provider's the model can fit), or the AI provider itself failed (rate limit, capacity, quota). The fix path depends on which bucket. Start with the text-extraction check — try selecting a sentence in the PDF — because most "PDF summary failed" reports turn out to be image-only scans that need OCR before any summarizer can work. This page walks through each cause with the specific fix.

Check these first

• The PDF is scanned image-only with no text layer — common for older documents and photographed pages.
• The PDF is password-protected, encrypted, or signed in a way that blocks text extraction.
• The PDF is partially loaded — Reduz tried to extract before Chrome finished rendering the document.
• The PDF is opened in an embedded viewer (some web apps wrap PDFs in custom viewers that block extension access).
• The PDF is very large (200+ pages) and exceeds the AI provider's the model can fit.
• Extraction succeeded but the AI provider failed (rate limit, capacity, quota) — a separate problem.

Fix it in this order

1. 1

   Try selecting text in the PDF

   Click and drag to select a sentence in the PDF. If text selection works, the PDF has a text layer and Reduz should be able to extract from it. If selection produces no highlight (or you see strange gibberish), the PDF is image-only or has a damaged text layer — run OCR first.

2. 2

   Run OCR on image-only PDFs

   For scanned PDFs without a text layer, run OCR before summarizing. Free options: native macOS Preview (Tools → Annotate), Adobe Acrobat (free OCR for one document), online OCR tools, or open-source tesseract. Save the OCRed version and reopen in Chrome.

3. 3

   Open the PDF in Chrome's built-in viewer

   Chrome's native PDF viewer is the most reliable surface for extension extraction. If the PDF is open in an embedded web viewer (Box, DocuSign, some SharePoint configurations), download it and reopen in Chrome's built-in viewer.

4. 4

   Wait for the PDF to fully load before clicking Reduz

   Large PDFs render progressively. If you click Reduz before all pages are rendered, extraction may catch only the first few pages. Wait until the page counter at the bottom shows the full page count, then click Reduz.

5. 5

   Summarize a page range for long PDFs

   For 200+ page PDFs that exceed the AI provider the model can fit, summarize a specific page range or chapter. Open the relevant pages, select the text you want summarized, then right-click "Summarize text with Reduz." Alternatively switch to a larger-context model — Claude Sonnet 4.6 or GPT-5.5 handle longer documents.

6. 6

   Check if the AI provider failed (not the extraction)

   If you saw extracted text on screen but the summary itself failed, the issue is the AI provider — rate limit, capacity, quota. Check the provider-specific troubleshooting pages or switch providers in Reduz settings.

7. 7

   Switch to a larger-context provider for long PDFs

   For 50-100 page PDFs that exceed smaller models, switch Reduz to Claude Sonnet 4.6 or GPT-5.5(longer the model can fits). For PDFs that exceed even those, page-range selection is the right approach.

Diagnosis

Extraction-bound

Image-only PDFs and PDFs with broken text layers need OCR. No summarizer can work with text it can't see.

Encryption-bound

Password-protected and encrypted PDFs block extension access to text content. Decrypt the PDF first (the legal way, using your password) and re-open.

Viewer-sensitive

Embedded viewers (Box, DocuSign, custom SharePoint PDFs) can block extension access. Download the PDF and open it in Chrome's built-in viewer for the most reliable extraction.

Size-bound

Long PDFs can succeed at extraction but fail at the AI provider when the token count exceeds the model's the model can fit. Switch to Claude Sonnet 4.6 or GPT-5.5, or summarize a page range.

Provider-bound

If extraction worked (you saw extracted text) but generation failed, the issue is the AI provider side — rate limit, capacity, or quota. Check provider-specific troubleshooting pages.

Smaller sections or larger-context providers

Most "PDF summary failed" reports fix with one of three moves: run OCR on image-only scans, summarize a page range instead of the whole document, or switch to a larger-context AI provider. Reduz keeps these moves all available — selected-text mode handles page ranges, provider switching takes one click in settings, and Hosted Free gives you a fallback that doesn't depend on your BYOK provider quotas. For sensitive PDFs (drafts, pre-prints, internal reports), Reduz reads the document from your Chrome tab without uploading the file anywhere — important when the document shouldn't live on a vendor server even after OCR or repair.

Frequently asked questions