> ## Documentation Index
> Fetch the complete documentation index at: https://docs.zerotwo.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Supported File Types

> Every file format ZeroTwo accepts, how it processes each one, and what you can do with it.

ZeroTwo supports a broad range of file types across documents, code, data, media, and notebooks. The processing method depends on the file category — here is a complete breakdown.

## Documents

ZeroTwo extracts full text and structure from all major document formats, enabling Q\&A, summarization, translation, and data extraction.

| Format       | Extensions                       | How ZeroTwo Processes It                                                     | Best Use Case                                   |
| ------------ | -------------------------------- | ---------------------------------------------------------------------------- | ----------------------------------------------- |
| PDF          | `.pdf`                           | Full text extraction; scanned PDFs use OCR                                   | Document Q\&A, summarization, compliance review |
| Word         | `.docx`, `.doc`                  | Text and heading structure parsed; tables preserved                          | Editing, summarizing, extracting key data       |
| PowerPoint   | `.pptx`, `.ppt`                  | Slide text extracted in order; speaker notes included                        | Presentation review, slide content extraction   |
| Excel        | `.xlsx`, `.xls`                  | Cell data extracted across all sheets; formula values (not strings) captured | Data analysis, table Q\&A, number extraction    |
| OpenDocument | `.odt`, `.ods`, `.odp`           | Text parsed from OpenDocument equivalents                                    | Document, spreadsheet, and presentation review  |
| Apple iWork  | `.pages`, `.numbers`, `.keynote` | Automatically converted and extracted                                        | Document, spreadsheet, and presentation review  |

**What you can ask:**

* "Summarize the executive summary of this report."
* "What are the main arguments in chapter 3?"
* "Extract all deadlines and owners from this contract."
* "Compare these two proposal versions and highlight what changed."
* "Translate this document from Spanish to English."

## Code Files

ZeroTwo reads the full source file with syntax awareness — it understands language structure, can follow function calls, identify patterns, and reason about logic.

| Language         | Extensions                 |
| ---------------- | -------------------------- |
| JavaScript / JSX | `.js`, `.jsx`              |
| TypeScript / TSX | `.ts`, `.tsx`              |
| Python           | `.py`                      |
| Java             | `.java`                    |
| C / C++          | `.c`, `.cpp`, `.h`, `.hpp` |
| C#               | `.cs`                      |
| Go               | `.go`                      |
| PHP              | `.php`                     |
| Ruby             | `.rb`                      |
| Swift            | `.swift`                   |
| Kotlin           | `.kt`                      |
| Scala            | `.scala`                   |
| Rust             | `.rs`                      |
| Shell / Bash     | `.sh`, `.bash`, `.zsh`     |
| PowerShell       | `.ps1`                     |
| SQL              | `.sql`                     |
| R                | `.r`, `.R`                 |
| Julia            | `.jl`                      |
| LaTeX            | `.tex`                     |

**What you can ask:**

* "Review this file for bugs and security vulnerabilities."
* "Explain what each function does."
* "Refactor this code to follow modern best practices."
* "Write unit tests for the functions in this file."
* "Convert this Python script to TypeScript."
* "Add type annotations and docstrings throughout."

## Data Formats

Structured data files are parsed and made available for analysis, statistics, schema description, and visualization prompts.

| Format  | Extensions      | Notes                                                               |
| ------- | --------------- | ------------------------------------------------------------------- |
| CSV     | `.csv`          | Most common tabular format; headers detected automatically          |
| TSV     | `.tsv`          | Tab-separated values; same processing as CSV                        |
| JSON    | `.json`         | Nested structures fully parsed; arrays of objects treated as tables |
| YAML    | `.yaml`, `.yml` | Configuration and structured data                                   |
| TOML    | `.toml`         | Configuration files                                                 |
| XML     | `.xml`          | Hierarchical data parsed; large files may be summarized             |
| Parquet | `.parquet`      | Columnar binary format; schema and column data extracted            |
| Feather | `.feather`      | Columnar binary format; schema and column data extracted            |
| Arrow   | `.arrow`        | Apache Arrow columnar format                                        |
| Avro    | `.avro`         | Schema-aware serialization; data and schema read together           |

**What you can ask:**

* "What are the column names and data types in this dataset?"
* "Which product had the highest sales in Q3?"
* "Describe the schema of this JSON file."
* "Find any rows with missing values."
* "Show the revenue trend by month as a bar chart."

## Markup Files

Markup and stylesheet files are read as text with structural awareness — ZeroTwo understands HTML semantics, CSS selectors, Markdown formatting, and GraphQL schema syntax.

| Format   | Extensions               | What ZeroTwo Can Do                                       |
| -------- | ------------------------ | --------------------------------------------------------- |
| HTML     | `.html`, `.htm`          | Convert to Markdown, review structure, extract content    |
| CSS      | `.css`, `.scss`, `.less` | Review for unused classes, explain selectors, refactor    |
| Markdown | `.md`, `.mdx`            | Rewrite with better structure, summarize, convert formats |
| GraphQL  | `.graphql`, `.gql`       | Explain schema, write queries, identify unused types      |

## Images

Images are analyzed by ZeroTwo's vision model, which can describe content, read text (OCR), identify objects, and answer specific visual questions.

**Supported extensions:** `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`

**What you can ask:**

* "What is in this image?"
* "Extract all the text from this screenshot."
* "Describe the chart in this image and summarize the data shown."
* "What does the error message in this screenshot say?"
* "Is there anything unusual or out of place in this photo?"
* "Identify all the UI elements in this wireframe."

<Note>
  Very small or very low-resolution images may produce less detailed analysis. For text extraction (OCR), use clear, high-contrast images at a minimum of 800×600 pixels for best results. Animated GIFs — only the first frame is analyzed.
</Note>

## Audio

Audio files are automatically transcribed using Whisper (OpenAI's speech recognition model). The transcript is then available for discussion, summarization, and analysis.

**Supported extensions:** `.mp3`, `.wav`, `.m4a`, `.aac`, `.ogg`, `.flac`, `.opus`

**What you can ask:**

* "Transcribe this meeting recording."
* "Summarize the key decisions from this audio."
* "Pull out all action items and who they were assigned to."
* "What topics were discussed in this conversation?"

| Format | Notes                                                            |
| ------ | ---------------------------------------------------------------- |
| MP3    | Most common compressed audio format                              |
| WAV    | Uncompressed; larger files, no quality loss                      |
| M4A    | Apple's AAC format; common for voice memos and iPhone recordings |
| AAC    | Compressed audio, good quality at smaller sizes                  |
| OGG    | Open format, common in web and game audio                        |
| FLAC   | Lossless compressed audio                                        |
| OPUS   | Modern codec optimized for speech quality                        |

<Tip>
  For non-English audio, specify the language in your prompt for better transcription accuracy: "Transcribe this audio — the speaker is speaking in Japanese."
</Tip>

## Video

Frames are extracted from the video at regular intervals, then ZeroTwo's vision model analyzes each frame. This allows scene description, object identification, and visual Q\&A across the video.

**Supported extensions:** `.mp4`, `.mov`, `.webm`, `.avi`, `.mkv`

**What you can ask:**

* "Describe what happens in this video, step by step."
* "What products are shown in this advertisement?"
* "Identify the steps demonstrated in this tutorial video."
* "At what point does the presenter show the dashboard?"

| Format | Notes                                                          |
| ------ | -------------------------------------------------------------- |
| MP4    | Most compatible and widely used video format                   |
| MOV    | Apple QuickTime; common for screen recordings and iPhone video |
| WEBM   | Web-optimized; common for browser-captured video               |
| AVI    | Older Microsoft format; still widely supported                 |
| MKV    | Matroska container; often used for high-quality video          |

<Warning>
  Video analysis is frame-based — fast-moving content between sampled frames may be missed. For analysis of specific moments, reference timestamps in your prompt: "At around the 2:30 mark, what is shown on screen?" Audio from videos is not automatically transcribed — upload the audio track separately for Whisper transcription.
</Warning>

**Maximum size:** 200 MB (Pro+), 50 MB (Free plan)

## Notebooks

Notebooks are read in full cell context — code cells, Markdown cells, and cell outputs (including printed results, dataframes, and errors) are all available for analysis.

**Supported formats:** `.ipynb` (Jupyter Notebook), `.Rmd` (R Markdown)

**What you can ask:**

* "Explain what this Jupyter notebook does, cell by cell."
* "Review the code in this notebook for errors and inefficiencies."
* "Summarize the findings from the output cells."
* "Convert this notebook's analysis steps into a clean Python script."
* "What statistical method is being used in this R Markdown document?"

## File Size Limits Summary

| File Category                                   | Max Size (Free) | Max Size (Pro+) |
| ----------------------------------------------- | --------------- | --------------- |
| Documents, Code, Data, Images, Audio, Notebooks | 50 MB           | 50 MB           |
| Video                                           | 50 MB           | 200 MB          |
| Files per message                               | 1               | 10              |

## Related Pages

<CardGroup cols={2}>
  <Card title="Upload Files" icon="upload" href="/tools/files/upload-files">
    How to attach files to chat messages.
  </Card>

  <Card title="Files Library" icon="folder" href="/tools/files/files-library">
    Browse and manage all your uploaded files.
  </Card>

  <Card title="Retention and Storage" icon="database" href="/tools/files/retention-and-storage">
    How files are stored and how long they persist.
  </Card>

  <Card title="Troubleshooting" icon="search" href="/tools/files/troubleshooting">
    Fix issues with unsupported formats and failed analysis.
  </Card>
</CardGroup>