Uploading Files & Document Analysis
Analyze anything you upload Attach any file to a chat — images, documents, spreadsheets, presentations, audio, video, code, archives, and more. There is no format allowlist: Krater accepts every fi...
Analyze anything you upload
Attach any file to a chat — images, documents, spreadsheets, presentations, audio, video, code, archives, and more. There is no format allowlist: Krater accepts every file type. Ask questions, get summaries, extract data, or feed the file into a slash command.
How to upload
- Click the + button in the chat input and choose Upload — or drag and drop files into the composer
- Select one or more files from your device
- Add your question or instructions
- Send your message
Limits:
- Any file type is accepted
- Maximum 500 MB per file
- Attach multiple files to a single message
How Krater reads your files
Common formats are handled natively: images go to vision models, PDFs and Office files are text-extracted, audio can be transcribed. Anything the model can’t read directly is staged into the agent’s sandboxed workspace and inspected with code (unzip, head, python, pandas and the like) — so even unusual formats still work.
Images
| Format | What you can do |
|---|---|
| JPG, PNG, WebP, GIF, HEIC, SVG | Ask questions, describe content, extract text (OCR) |
With /image |
Edit, transform, or remix uploaded photos |
Example prompts:
- “What’s in this image?”
- “Extract all the text from this screenshot”
- “What’s the color palette used here?”
- “Find any errors or issues in this UI screenshot”
Documents
| Format | Description |
|---|---|
| Multi-page documents, text extraction, table parsing | |
| DOCX / DOC | Microsoft Word, current and legacy |
| TXT, MD, RTF, EPUB, ODT | Plain text, Markdown, ebooks, and OpenDocument |
Example prompts:
- “Summarize this document in 3 bullet points”
- “What are the key findings in this report?”
- “Extract the main dates and deadlines mentioned”
Spreadsheets
| Format | Description |
|---|---|
| XLSX / XLS | Microsoft Excel (all sheets parsed) |
| CSV | Comma-separated values |
Example prompts:
- “Analyze this sales data and find trends”
- “Calculate the total from column B”
- “Create a summary of this spreadsheet”
Presentations
| Format | Description |
|---|---|
| PPTX / PPT | Microsoft PowerPoint (text from all slides) |
Example prompts:
- “Summarize this presentation”
- “What are the main points from each slide?”
- “Identify the key takeaways”
Audio & video
| Format | What you can do |
|---|---|
| MP3, WAV, M4A, OGG, FLAC | Transcription and analysis — /transcribe gives a verbatim transcript |
| MP4, MOV, MKV | Reference footage for /video generations and edits |
Example prompts:
- “Transcribe this audio”
- “Summarize what was discussed”
- “Extract action items from this meeting”
Everything else
Code files, JSON, logs, ZIP archives, research data (.r, .tex, .ipynb), and formats not listed here are all accepted — the agent inspects them in its sandbox.
Chat attachments vs Keep
Files you attach stay with the conversation — they are not added to your Keep file storage automatically. Keep is your permanent, organized storage (25–250 GB depending on plan). Generated outputs save to Keep, and you can save any AI response or file from the message actions. To keep an uploaded file long-term, organized, or reusable across chats, save it to Keep deliberately.
Tips for better results
Be specific
| Instead of… | Try… |
|---|---|
| “Tell me about this” | “What’s the main topic of this document?” |
| “Look at this” | “Are there any errors in this code screenshot?” |
| “Summarize” | “Give me a 3-sentence summary for executives” |
Provide context
Adding context helps the AI understand what you need:
- “This is a contract — highlight any unusual clauses”
- “This spreadsheet tracks Q3 sales — find the top performers”
- “This is a screenshot of my app’s error — help me debug”
Chain with commands
Upload a file, then use slash commands:
- Upload a photo → use
/imageto edit it - Upload a document → ask questions →
/summarizefor a condensed version - Upload audio →
/transcribe→ ask for key points
Document processing details
How it works
When you upload a document:
- Text extraction — All readable text is parsed
- Structure preservation — Tables, headings, and formatting are maintained
- Multi-page support — Each page/sheet/slide is processed
- OCR — Scanned documents and images with text are recognized
What’s included
| Document type | Extracted content |
|---|---|
| Text, tables, page structure | |
| DOCX | Full text content |
| XLSX | All sheets as structured data |
| PPTX | Text from all slides |
Privacy & security
- Attachments are uploaded over a secure connection and tied to your conversation
- They are not added to Keep automatically — you choose what to save
- Your content is not used to train AI models
Common use cases
Research:
- Summarize academic papers
- Extract citations and references
- Compare multiple documents
Work:
- Analyze reports and presentations
- Extract data from spreadsheets
- Review contracts and legal documents
Creative:
- Describe images for alt text
- Get feedback on designs
- Transform uploaded images with
/image
Learning:
- Explain complex documents
- Create study notes from textbooks
- Translate document content