Frequently Asked Questions

Detailed answers to the questions we hear most often — covering Gemini API setup, file support, OCR accuracy, privacy, language options, and export formats. If your question isn't here, email contact@notebooklm-editor.online.

Setup and Getting Started

What is Notebook LM Slide Editor?

Notebook LM Slide Editor is a free, browser-based tool for replacing text inside PDF documents and image-based presentation slides. It uses Google Gemini's multimodal AI to automatically detect the font family, size, weight, color, and spacing of any text region you select — then lets you paint matching replacement text directly over the original document. Nothing is installed on your computer, and no account is required. See the Getting Started Tutorial for a step-by-step walkthrough.

Why do I need a Gemini API key?

The AI text analysis feature (which detects font and color information) uses Google's Gemini multimodal model, accessed through the Google AI API. This API requires an authentication key. We do not provide a shared key because doing so would pool all users under a single rate limit, causing failures during peak usage. Your own free key gives you a personal quota that is more than sufficient for regular editing work.

How do I get a free Gemini API key?

Go to aistudio.google.com and sign in with any Google account. Click "Get API key" → "Create API key" → "Create API key in new project." Copy the key and paste it into the Settings field in the editor (click the gear icon in the top navigation). The process takes less than 2 minutes and requires no credit card.

Is there a cost to use Notebook LM Slide Editor?

The editor itself is completely free — no subscription, no one-time fee, no account required. The Gemini API has a generous free tier that covers normal editing usage for individuals and small teams. If you perform a very high volume of OCR analyses (thousands per day), you may eventually exceed the free quota. For typical use cases — occasional to moderate document editing — you will never reach a paid threshold.

Can I use the editor without the Gemini API key?

Yes, for manual edits. You can upload a document, draw selection boxes, manually enter font and color settings in the sidebar, type replacement text, and export the result — all without an API key. The only feature that requires the API key is the "Run AI OCR" button, which automatically detects the style parameters. Manual editing is slower but fully functional.

File Formats and Upload

What file types does the editor accept?

The editor accepts PDF (including multi-page PDFs), PNG, JPG/JPEG, and WebP files. PDFs are automatically rendered into page images on upload. For best OCR results, use source files at 300 DPI or higher. When exporting from presentation software, use the highest-quality PDF option available.

Is there a file size limit?

There is no hard file size limit enforced by the editor. However, because all rendering happens locally in your browser, very large files — PDFs over 100 pages or images over 20MB — may load slowly on older or memory-constrained devices. If performance is an issue, split large PDFs into smaller chunks (30–50 pages) using any free PDF-splitting tool before uploading.

Can I edit a password-protected PDF?

Not directly — password-protected PDFs cannot be rendered without the password being removed first. Remove the protection using a PDF management tool before uploading. (Adobe Acrobat, Smallpdf, and PDF24 all offer this capability.)

Can I upload multiple documents at the same time?

The editor currently processes one document at a time. Complete your edits on the first document, download the result, and then upload the next. Batch processing across multiple documents is on the roadmap for a future update.

What happens to my file after I upload it?

Your file is loaded into your browser's local memory only — it is never uploaded to our servers. The document exists in RAM for the duration of your editing session and is released when you close the tab. We have no copy of your file and no record that you uploaded it.

OCR Accuracy and Limitations

How accurate is the AI font and color detection?

For standard printed text in common fonts (Arial, Helvetica, Inter, Roboto, Open Sans, Times New Roman, etc.) at 200 DPI or higher, the AI achieves high accuracy on font family, weight, size, and color. Accuracy is lower for: text smaller than 10pt at screen resolution, decorative or display-only fonts, text on highly patterned or photographic backgrounds, and handwritten content.

What if the AI detects the wrong font family?

The AI's font suggestion is a starting point. After OCR runs, click the Font Family dropdown in the sidebar and select a different font. The live canvas preview updates instantly so you can compare visually until you find the closest match. Many custom or corporate fonts have a near-equivalent available in Google Fonts — trying the 2–3 closest alternatives usually gets you there.

What if the detected color is slightly off?

Click the color swatch next to the Text Color or Background Color fields to open a color picker and manually adjust the hex value. For precise matching, use your browser's built-in eyedropper tool (available in Chrome and Edge) to sample the exact color from the original text, then paste that hex code into the color field.

Does OCR work for handwritten text?

The Gemini model can read some handwritten text, but accuracy varies significantly by handwriting style. More importantly, "font detection" doesn't apply to handwriting — there's no font family or weight to detect. For handwritten documents, skip the OCR step and manually configure the text style in the sidebar to approximate the handwriting style you want to reproduce.

Can the editor detect text in tables or complex layouts?

Yes, but accuracy depends on the complexity of the layout. For text in simple table cells with clear borders and solid backgrounds, OCR works well. For text embedded in complex charts, overlapping elements, or dense multi-column layouts, you may need to use smaller selection boxes that isolate individual text elements, and verify the detected values manually before applying.

Privacy and Data Security

Where are my documents stored?

They are not stored anywhere outside your browser. Documents are loaded into browser memory (RAM) for the duration of your editing session. When you close the tab, the document is released from memory. We operate no document database and have no access to files you upload.

What data is sent to Google Gemini?

Only the small cropped image region you explicitly select for OCR analysis is sent to the Gemini API — not the entire page, and not the full document. This cropped region is typically 50–500 pixels across, depending on how large your selection box is. It is transmitted over HTTPS and discarded by the API once the style response is returned. Google's API Terms of Service govern how Gemini handles this data.

Does the editor use cookies or analytics?

The editor uses no cookies for user tracking. UI language preference and editor settings are saved in your browser's localStorage — they stay on your device only, are never transmitted to us, and are deleted if you clear your browser data. The site includes Google AdSense advertising scripts, which may use cookies for ad personalization as described in the Privacy Policy.

Is the Gemini API connection encrypted?

Yes. All API calls are made over HTTPS (TLS 1.2 or 1.3). Your Gemini API key and the image region are transmitted securely. The API key itself is stored only in your browser's memory during the session — it is not saved to any server.

Supported Languages

Which UI languages are available in the editor?

The editor interface is available in 15+ languages: English, Korean (한국어), Japanese (日本語), Chinese Simplified (简体中文), Chinese Traditional (繁體中文), Spanish (Español), French (Français), German (Deutsch), Italian (Italiano), Portuguese (Português), Russian (Русский), Arabic (العربية), Hindi (हिन्दी), Vietnamese (Tiếng Việt), and Thai (ภาษาไทย). Use the globe icon in the top navigation to switch languages.

Can I edit documents written in non-English languages?

Yes. The Gemini OCR engine recognizes text in all the UI languages listed above, plus many others. You can replace text in any language that the substituted font supports — you are not limited to the languages in the UI selector. The editor is particularly well-tested for Korean, Japanese, and Chinese documents because those are the most common use cases for NotebookLM exports.

Does the editor automatically translate text?

No. The editor replaces text — it does not generate translations. You provide the translated text yourself (or copy it from a translation tool). The editor's role is to apply the replacement text in a style that matches the original document, so the edit looks seamless rather than obviously inserted.

Export and Output

What export formats are available?

You can export as a multi-page PDF (recommended for sharing and printing) or as individual PNG images per page (useful when you need to insert a specific edited slide into another application). Both formats include all text overlays baked into the output.

Does the exported PDF match the original quality?

The output quality depends on the resolution of the source file. The background (original document) is rendered at the source file's resolution. Overlaid text is rendered at a high raster resolution that is indistinguishable from the original at normal viewing and print sizes. Close-up inspection of the text overlay area may show minor rasterization at very high magnification, but this is invisible in practical use.

Can I save my editing session and continue later?

Not currently. When you close the tab, the session ends and cannot be resumed. For long editing projects, we recommend periodically downloading a work-in-progress PDF as a checkpoint — even before all edits are complete — so you have a saved intermediate state if the session is interrupted.

Can I undo an edit?

Yes. The editor has an Undo button (the curved arrow icon in the top navigation) that removes the most recently applied text overlay. You can undo multiple overlays in sequence. However, undo history is cleared when you close the tab — it does not persist between sessions.

NotebookLM Integration

What is the connection between this tool and Google NotebookLM?

Notebook LM Slide Editor was originally designed to complement Google NotebookLM. NotebookLM generates AI-powered study guides, summaries, and briefing documents from your uploaded source materials, and exports them as PDFs. These exports are often 90% ready to use but need small corrections — a heading that doesn't quite fit the audience, a phrase that's slightly off, a statistic that needs updating. This editor lets you make those corrections quickly while preserving the original visual design of the NotebookLM output.

Can I use the editor for documents other than NotebookLM exports?

Absolutely. The editor works with any PDF or image file regardless of its source. Common uses include localizing corporate presentations for new markets, correcting published PDF reports, updating slide decks for new quarters or campaigns, and adapting educational materials for specific audiences. The NotebookLM integration is the origin story, but it's far from the only use case.