--- title: "AI Activate - Supported File Types Reference" slug: "ai-activate-supported-file-types" description: "Explore how AI Activate categorizes file types for AI access, detailing Content Activated and Metadata Only files for optimal information retrieval." tags: ["AI Activate", "Best Practices"] updated: 2026-10-02T20:17:54Z published: 2026-10-02T20:17:54Z canonical: "docs.nasuni.com/ai-activate-supported-file-types" --- > ## Documentation Index > Fetch the complete documentation index at: https://docs.nasuni.com/llms.txt > Use this file to discover all available pages before exploring further. # AI Activate - Supported File Types Reference The following is the authoritative list of file types that AI Activate can activate for AI access, along with what activation does and does not cover. Nasuni updates this list over time as support expands. ## How AI Activate Handles Your Files AI Activate activates your file content so that AI tools can find and use the right information when you ask a question. Every file in scope falls into one of two categories: - **Content Activated.** AI Activate extracts and activates the text inside the file, so that AI tools can find the file by the file's contents, not just by the file's name. Activated text makes a file usable as grounding for an AI answer. - **Metadata Only.** AI Activate records the file's name, path, type, size, and dates. AI Activate does not read the file’s contents. AI Activate records the file's existence, but AI tools cannot find the file by the file’s contents. A file’s format determines the file’s category, as described below. ## Content Activated File Types AI Activate reads and activates the text content of the following file types. Content in these formats is available to AI tools through AI Activate. - **Microsoft Office:** .doc, .docx, .docm, .dot, .dotm, .xls, .xlsx, .xlsm, .xlsb, .ppt, .pptx, .pptm, .pot - **OpenDocument:** .odt, .ods, .odp, .odf - **Apple iWork:** .pages, .numbers, .key - **PDF:** .pdf (text-based PDFs only, see the note on scanned PDFs below) - **Web and markup:** .html, .htm, .xhtml, .xml - **Email:** .eml, .msg - **eBook:** .epub - **Rich text:** .rtf - **Plain text and data:** .txt, .md, .markdown, .csv, .tsv - **Source code:** .java, .py, .c, .cpp, .js These formats cover most sources of business knowledge. AI Activate activates content in files up to 100 MB. AI Activate treats files larger than 100 MB as Metadata Only Files. AI Activate captures up to 50,000 characters (about 10,000 English words) from a Content Activated file. ## Metadata Only Files If a file is in a format that is not on the list above, or if activation is not possible for a file, AI Activate records only the file's metadata. Because the file content is not read, an AI tool cannot search the file based on its contents. This is an important distinction: Metadata Only files are not invisible. The file is still part of your scoped content and can still be discovered through AI Activate by the file’s name, path, and other metadata attributes. After a file has been identified that way, your AI tool can use its presence as a point of reference. In short: - **Content Activated File:** AI tools can find the file by the file's contents and use the file's text to ground answers. - **Metadata Only File:** AI tools can find the file by name, path, and other metadata attributes. ## What AI Activate Does Not Extract from Files AI Activate works from the text that already exists in a file. AI Activate does not generate context from images or non-text content types. Specifically, AI Activate does not support the following: - **Optical character recognition (OCR).** AI Activate does not convert images of text into searchable text. - **Image tag recognition.** AI Activate does not identify or label the contents of photos or graphics. - **EXIF data extraction.** AI Activate does not read embedded image metadata such as camera or location data. ### Example of How AI Activate Handles Scanned PDFs and Images For example, the .pdf format is on the Content Activated File Types list, but AI Activate activates only PDFs with real, selectable text. A scanned PDF is an image of a page rather than actual text. AI Activate treats a scanned PDF as Metadata Only: AI Activate records the file but cannot read the file's content. AI Activate also does not extract information contained in images, including: - Text embedded in a chart, diagram, screenshot, or photo inside a document. - Image-only PDFs and faxed or photographed documents. - Standalone image files such as .jpg, .png, .gif, .tif, and .heic. ## Is My Content Activated? - The file is a Word, Excel, PowerPoint, text, or text-based PDF document: yes, AI Activate activates the file’s content for AI access. - The file is a scanned PDF or a photo of a document: no, AI Activate records only the file’s name and metadata. - The information I need is inside a chart or image within a document: no, AI Activate does not read the information. - The file is an unusual or proprietary format not on the list: AI Activate treats the file as Metadata Only.