# Document Extractor

> >

- **Type:** Skill
- **Install:** `agentstack add skill-lirrensi-agent-sommelier-document-extractor`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [lirrensi](https://agentstack.voostack.com/s/lirrensi)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [lirrensi](https://github.com/lirrensi)
- **Source:** https://github.com/lirrensi/agent-sommelier/tree/main/skills/document-extractor

## Install

```sh
agentstack add skill-lirrensi-agent-sommelier-document-extractor
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# DocumentExtractor

Use `markitdown` to convert supported inputs into Markdown that is easier to inspect, search, and feed into LLM workflows.

## Check And Install

Check whether the CLI already exists:

```powershell
Get-Command markitdown -ErrorAction SilentlyContinue
markitdown --version
```

Install with `uv` first:

```powershell
uv tool install markitdown
uv tool install 'markitdown[pdf,docx,pptx]'
uv tool install 'markitdown[all]'
```

Fall back to `pipx` if `uv` is unavailable:

```powershell
pipx install markitdown
pipx install 'markitdown[pdf,docx,pptx]'
pipx install 'markitdown[all]'
```

If the tool is installed without the feature group you need, reinstall it with the narrower or broader extras set you actually want.

Read [references/feature-groups.md](references/feature-groups.md) before installing extras when the user only needs a subset of formats.

## Use The CLI

Convert a file and print Markdown to stdout:

```powershell
markitdown .\report.pdf
```

Write to a file:

```powershell
markitdown .\report.pdf -o .\report.md
```

Pipe binary input and give MarkItDown an extension hint:

```powershell
Get-Content .\report.pdf -AsByteStream | markitdown -x .pdf -o .\report.md
```

Use MIME or charset hints when the input source is ambiguous:

```powershell
markitdown -x .html -m text/html .\page.bin
markitdown -x .csv -c utf-8 .\data.txt
```

Keep inline `data:` URIs instead of truncating them:

```powershell
markitdown --keep-data-uris .\page.html -o .\page.md
```

## Choose Extras Deliberately

- Install base `markitdown` for lightweight text-like inputs and general conversion.
- Install targeted extras when the user only needs specific formats.
- Install `[all]` only when broad coverage matters more than dependency size.
- Reinstall with `az-doc-intel` when using Azure Document Intelligence.
- Reinstall with `audio-transcription` or `youtube-transcription` only for transcription workflows.

The current extras list and format mapping lives in [references/feature-groups.md](references/feature-groups.md).

## Common Workflows

Convert Office or PDF documents:

```powershell
markitdown .\slides.pptx -o .\slides.md
markitdown .\notes.docx -o .\notes.md
markitdown .\table.xlsx -o .\table.md
markitdown .\scan.pdf -o .\scan.md
```

Convert a YouTube URL or archive when the relevant support is installed:

```powershell
markitdown "https://www.youtube.com/watch?v=dQw4w9WgXcQ" -o .\video.md
markitdown .\bundle.zip -o .\bundle.md
```

Use Azure Document Intelligence for extraction:

```powershell
markitdown .\scan.pdf -d -e "https://.cognitiveservices.azure.com/" -o .\scan.md
```

List installed third-party plugins:

```powershell
markitdown --list-plugins
markitdown --use-plugins .\input.pdf -o .\input.md
```

## Use The Python API

Use the Python API when the user needs MarkItDown inside a script instead of as a standalone command:

```python
from markitdown import MarkItDown

md = MarkItDown(enable_plugins=False)
result = md.convert("report.pdf")
print(result.markdown)
```

Use a configured endpoint for Document Intelligence:

```python
from markitdown import MarkItDown

md = MarkItDown(docintel_endpoint="https://.cognitiveservices.azure.com/")
result = md.convert("scan.pdf")
print(result.markdown)
```

## Troubleshoot Quickly

- If `markitdown` is missing, install it with `uv tool install ...` or `pipx install ...`.
- If a format is unsupported, check whether the right extra was installed first.
- If stdin conversion looks wrong, add `-x`, `-m`, or `-c` hints.
- If `-d` fails, verify the endpoint and that `az-doc-intel` support is installed.
- If plugin behavior is expected, run `markitdown --list-plugins` and then add `--use-plugins`.
- If output details are unclear, run `markitdown --help` and then check the upstream docs.

## Last Resort

Use these sources when local behavior is unclear or the package changes:

- CLI help: `markitdown --help`
- Main docs: `https://github.com/microsoft/markitdown/tree/main`
- README: `https://raw.githubusercontent.com/microsoft/markitdown/main/README.md`

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [lirrensi](https://github.com/lirrensi)
- **Source:** [lirrensi/agent-sommelier](https://github.com/lirrensi/agent-sommelier)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-lirrensi-agent-sommelier-document-extractor
- Seller: https://agentstack.voostack.com/s/lirrensi
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.