Skip to main content
ZeroTwo supports attaching files directly to chat messages. The AI analyzes the file content and incorporates it into its response — enabling document Q&A, code review, image analysis, data exploration, audio transcription, and more, all within a single conversation.

​
How to Attach Files

There are two ways to attach files to a message: Paperclip icon: Click the paperclip icon in the prompt bar to open the system file picker. Select one or more files from your device. Selected files appear as preview thumbnails or document cards above the text input field. Drag and drop: Drag files from your computer directly onto the chat area and release. They are queued as attachments for your next message and appear as previews in the prompt bar. You can remove any attachment before sending by clicking the × on its preview card.

​
Attachment Limits

LimitFree PlanPro and Above
Files per message1Up to 10
Max file size per file50 MB50 MB
Max video file sizeNot supported200 MB
Upload timeout60 seconds60 seconds

​
What ZeroTwo Does With Each File Type

ZeroTwo extracts full text content from all pages of the PDF, preserving page structure, headings, and paragraph flow where possible.Best uses:
  • Q&A on specific sections: “What are the payment terms in Section 4 of this contract?”
  • Summarization: “Summarize the executive summary of this report.”
  • Table extraction: “Extract all the pricing data from this document as a table.”
  • Compliance review: “Identify any clauses that could be problematic for a GDPR-compliant business.”
  • Comparison: Upload two PDFs and ask “What are the main differences between these two proposals?”
Be specific in your prompt. “Summarize only the executive summary section” gets better results than “summarize this document.” The more focused the question, the more useful the response.
ZeroTwo parses the full text and structure of Word documents — headings, paragraphs, tables, and lists — making the content available for analysis and editing.For PowerPoint files, ZeroTwo extracts slide titles, body text, and speaker notes from each slide.Best uses for DOCX:
  • Edit or rewrite sections: “Rewrite the introduction to be more concise.”
  • Extract specific information: “List all action items mentioned in this meeting notes document.”
  • Summarize or restructure: “Convert this report into a bulleted executive summary.”
Best uses for PPTX:
  • Summarize slides: “What are the main points of this presentation?”
  • Generate speaker notes: “Write speaker notes for each slide.”
  • Extract data: “List all the statistics mentioned across these slides.”
ZeroTwo reads spreadsheet data as structured tabular content. For Excel files, it parses cell values across sheets. For CSV files, it reads rows and columns directly.Best uses:
  • Statistical analysis: “What is the average monthly revenue across these quarters?”
  • Trend identification: “Is there a correlation between column A and column C?”
  • Summary statistics: “Give me the min, max, and average for each numeric column.”
  • Insight generation: “What are the top 5 customers by total spend?”
  • Formula suggestions: “Suggest formulas to calculate a 30-day rolling average for this data.”
  • Chart description: “Describe what a bar chart of this data would look like.”
For complex multi-sheet Excel files, specify which sheet you want analyzed: “Using the ‘Q3 Sales’ sheet, identify the top performing regions.”
ZeroTwo passes images to a vision-capable model for analysis. The model can describe what it sees, read text within the image (OCR), identify objects and people, analyze charts and graphs, and answer visual questions.Best uses:
  • Describe images: “Describe this photo in detail.”
  • Extract text (OCR): “Transcribe all text visible in this screenshot.”
  • Analyze charts: “What trend does this graph show? Summarize the key data points.”
  • Review UI screenshots: “What issues do you notice with this UI layout?”
  • Accessibility: “Write an alt-text description for this image.”
  • Identify content: “What products are shown in this product photo?”
Image analysis requires a vision-capable model. If your currently selected model does not support vision, ZeroTwo will prompt you to switch to a compatible one. Check the capabilities indicator in the model picker to confirm vision support before attaching an image.
ZeroTwo automatically transcribes audio files using OpenAI’s Whisper model, which supports dozens of languages. The transcript is then passed to your selected chat model as text, making the spoken content available for any text-based analysis.Transcription is fully automatic — there are no extra steps. The transcript appears in the chat and becomes part of the conversation context.Best uses:
  • Meeting recordings: “Summarize the key decisions made in this recording.”
  • Interview transcripts: “Extract all questions asked in this interview.”
  • Lecture notes: “Create structured notes from this lecture recording.”
  • Podcast summaries: “What are the main topics discussed in this episode?”
  • Action items: “List all action items and who they were assigned to.”
Whisper handles accented speech, background noise, and multiple speakers reasonably well. For very long audio files (over 30 minutes), consider splitting them into segments for best results.
ZeroTwo extracts representative frames from the video at intervals and passes them to a vision-capable model for analysis. This enables description of video content, extraction of visible text, and analysis of what appears on screen.Video processing runs as a background task. Results are returned to the chat when processing is complete.Best uses:
  • Screen recordings: “Describe what happens in this screen recording step by step.”
  • Video presentations: “What slides or content are shown in this video walkthrough?”
  • Visual content review: “What products or environments are shown in this footage?”
  • Text extraction: “Extract any text visible on screen throughout this recording.”
Video files are capped at 200 MB. ZeroTwo samples frames rather than processing every frame, so fine-grained second-by-second analysis is not possible for long recordings. For detailed analysis, consider shorter clips or specific time-range exports.
ZeroTwo reads source code files as text and passes the full file content to the model. This gives the AI complete visibility into your code for analysis, review, refactoring, and documentation tasks.Best uses:
  • Code review: “Review this file for security vulnerabilities and code quality issues.”
  • Explanation: “Explain what this function does, line by line.”
  • Refactoring: “Refactor this to use async/await instead of callbacks.”
  • Test generation: “Write unit tests for the functions in this file.”
  • Documentation: “Generate JSDoc comments for all exported functions.”
  • Bug identification: “Find any potential null pointer exceptions in this code.”
When attaching code files, always include a clear, specific instruction in your message. “Review this for security issues” gets more useful results than “look at this code.” You can attach multiple related files together on Pro+ plans for cross-file analysis.
ZeroTwo parses structured data files and makes their content available for analysis, transformation, and query generation.Best uses for JSON:
  • Explore API responses: “Describe the structure of this JSON and explain what each field means.”
  • Transform data: “Convert this JSON array into a CSV format.”
  • Validate structure: “Does this JSON match the expected schema for a Stripe webhook event?”
Best uses for YAML:
  • Review configuration files: “Are there any misconfigurations in this Kubernetes deployment YAML?”
  • Explain settings: “Explain each field in this GitHub Actions workflow file.”
Best uses for Parquet/CSV:
  • Statistical analysis, summarization, insight generation (same as Excel/CSV — see above)
ZeroTwo reads all cells in a Jupyter Notebook — code cells, markdown cells, and cell outputs — making the full notebook content available for analysis and explanation.Best uses:
  • Explain ML experiments: “Explain what this notebook is doing and what the results show.”
  • Review code: “Are there any potential issues with the data preprocessing steps in this notebook?”
  • Summarize findings: “Summarize the conclusions drawn from this analysis.”
  • Generate documentation: “Write a README description for this notebook explaining its purpose and methodology.”

​
Multiple Files in One Message

On Pro and above plans, you can attach up to 10 files in a single message. This enables cross-document analysis:
  • Compare two contract PDFs side by side: “What are the key differences between these two agreements?”
  • Analyze a CSV and a PDF together: “Does the data in this spreadsheet match the targets described in this report?”
  • Review a set of related code files: “Review these three files together and identify any inconsistencies in how errors are handled.”
  • Combine an audio transcript with a document: “Does the meeting recording align with the agenda document?”
When attaching multiple files, ZeroTwo processes all of them together and presents them to the model as a unified context payload.

​
Files Library

All files you upload are stored in your Files Library at /files. From there you can:
  • Browse all previously uploaded files with metadata (name, size, upload date)
  • Re-attach files from the library to new messages without re-uploading
  • Delete files you no longer need
Files uploaded in chat are associated with your account and stored on ZeroTwo’s servers. Review your Files Library periodically and delete any sensitive or no-longer-needed files.

​
Related Pages