Chat with PDF
Attach a PDF, Word file, spreadsheet or slide deck and ask any model about it. Long files are fine: the model reads the parts it needs. And when the PDF itself needs changing, ask for that too.
Summarise with Gemini, question with Claude, double-check with GPT. Same file, same thread.
Big files are read in parts; scanned pages are read as images by vision models.
Replace text, rotate, reorder or delete pages, merge, watermark, number pages, fill forms.
PDF, DOCX, XLSX, PPTX, the OpenDocument formats, text and data files like CSV, JSON and Markdown, and source code. Up to 25 MB per file. Spreadsheets keep their sheet structure, so questions like "which region grew fastest?" work.
Ask in plain words: "remove pages 4 to 6", "merge these two", "add page numbers", "fill this form with my details". The edited copy appears as a download card and the original stays untouched. The model can also keep editing files it made earlier in the chat.
File uploads come with Pro ($20 a month, with a 7-day free trial on monthly billing) and above. Plus and Ultra raise the upload limits.
Yes. When a page has no text layer, vision models read it as an image, so scans and photographed documents work.
Gemini and Claude are popular picks for long documents. On Portail you can try both on the same file in one thread.
No. Edits produce a new copy you download. The file you uploaded stays as it was.
Every model, one thread, one memory. Free to start.