Multiple files. One conversation.
Attach up to 8 files to a conversation. Choose files from your library and continue a chat or start a new one.
Learn how to summarize Word documents and PDFs, turn source material into slide outlines, and clean Excel tables. Find upload limits, example prompts and export formats below.
Attach up to 8 files to a conversation. Choose files from your library and continue a chat or start a new one.
Review the document references alongside an answer, then check important conclusions and figures against the originals.
Upload DOC and PPT files for automatic conversion before extraction, without saving a new format on your computer.
Edit text results and download documents, export outlines to PPTX, and download processed tables as XLSX.
Upload in a studio, or select files in the library and add them to the current chat or a new chat.
DocDesk extracts readable text and tables. Older DOC and PPT files convert first; wait until processing finishes.
Ask for a summary, a slide outline, or table statistics. State the format you want, then follow up with “Make it shorter” or “table statistics”.
Check the sources and numbers, then edit or download the result. Treat generated content as a working draft to review.
Copy an example, add your files in a studio, then send it. Replace keywords such as “project budget” to match your sources.
Summarize the key points and conclusions of this document.
Find passages about the project budget in this PDF and list the sources I can check.
Create a presentation outline from these documents, with a title and key points for each slide.
Clean tables: trim whitespace, remove empty rows, deduplicate entire rows, and show table statistics.
Extract text and native tables; create editable Word content.
Original layout, images and tracked changes are not fully preserved.
Extract readable text by page for summaries, retrieval and questions.
Run OCR on scans and decrypt protected files first. PDF editing and export are not available.
Read slide text, tables and speaker notes; create outlines and PPTX files.
New decks use simple text layouts, without recreating the original slide design.
Read worksheets, trim whitespace, remove empty rows, deduplicate whole rows and summarize numbers.
Formulas use cached values; no recalculation or macros. Save XLS as XLSX before uploading.
Convert to DOCX or PPTX on the server, then continue with the same document workflow.
Up to 10 MB per file. Originals are retained; complex layouts may change. Failed conversions can be retried.
Read UTF-8 text as reference material, a summary source or a starting draft.
DOCX, PDF, PPTX, XLSX, CSV, MD and TXT support up to 20 MB per file.
Extract source sentences, build structured template drafts, retrieve passages by keyword, and run defined table cleanup and statistics without a model. This mode does not provide semantic rewriting or open-ended Q&A.
With a model configured, use selected sources for semantic questions, rewriting and content generation. “Configured” describes server settings, not a verified connection; responses depend on model service availability.
Table calculations follow implemented rules. Configuring a model does not add arbitrary formulas, conditional filters or code execution.
Attach up to eight files to a conversation. DOCX, PDF, PPTX, XLSX, CSV, Markdown and plain text files can be up to 20 MB each. Older DOC and PPT files can be up to 10 MB each and are converted automatically. Save older XLS spreadsheets as XLSX before uploading.
DocDesk reads the text layer of a PDF. For an image-only scan, run optical character recognition (OCR) first and upload the resulting searchable PDF. Decrypt password-protected PDFs before upload. Review the extracted text, especially on pages with complex layouts.
Yes. Open the answer in your conversation. Download Word saves a text result as DOCX, Download PPT saves an outline as PPTX, and Download Excel saves the result tables as XLSX. More also offers a Markdown download of the response.
DocDesk keeps the original DOC or PPT file and uses the converted copy to extract content. Complex layouts and embedded objects may change during conversion. If conversion fails, retry from the file’s processing status or upload a copy saved in Office as DOCX or PPTX.
Uploaded files and saved results belong to your account. Chat history is stored in the browser where you started the chat. Clearing browser data can remove it. In AI mode, relevant text from selected files is sent to the configured model service for processing.
DocDesk independently implements its own document processing tools. Anthropic’s publicly shared Claude Skills document workflows informed their design. DocDesk does not run the repository’s skill scripts and is not an official or endorsed Anthropic product.