# Convert PowerPoint (.pptx) and Excel (.xlsx) to Markdown for LLMs

> Turn slide decks and spreadsheets into clean Markdown with tables for RAG and agents with the Reducto Parse API.

Canonical: https://reducto.ai/cookbooks/parse-pptx-xlsx-for-llms

Send a .pptx or .xlsx to Reducto Parse and get Markdown back: slide titles as headings, bullet text, and tables as Markdown or HTML tables, with layout blocks and page numbers for citations. It is the same call you use for PDFs, so one pipeline covers decks, spreadsheets and reports.

## What you'll build

Synthetic 3-slide board deck (or your own file)

- Upload the .pptx or .xlsx, or pass a URL.
- Call /parse; Office files are converted and parsed like any document.
- Read result.chunks[].content as Markdown.
- Use blocks[] page numbers to cite slides in answers.

## Setup

Grab an API key from <a href="https://studio.reducto.ai" rel="noopener noreferrer" target="_blank">studio.reducto.ai</a> -> API Keys -> Create new API key, then set it and install the SDK:

```bash
export REDUCTO_API_KEY="your-api-key-here"
pip install reductoai   # or: npm install reductoai
```

## Step 1: Call `/parse`

The options that matter for this job:

- `settings.model`: Reads slides as rendered, including tables and charts.
- `formatting.table_output_format`: Tables as Markdown for LLM context.

```python
# pip install reductoai
from pathlib import Path
from reducto import Reducto

client = Reducto()  # reads REDUCTO_API_KEY

upload = client.upload(file=Path("deck.pptx"))
result = client.parse.run(
    input=upload.file_id,
    settings={
        "model": "r-1",
    },
    formatting={
        "table_output_format": "md",
    },
)
print(result)
```

```javascript
// npm install reductoai

const client = new Reducto(); // reads REDUCTO_API_KEY

const upload = await client.upload({ file: fs.createReadStream("deck.pptx") });
const result = await client.parse.run({
  input: upload.file_id,
  settings: {
    model: "r-1",
  },
  formatting: {
    table_output_format: "md",
  },
});
console.log(JSON.stringify(result, null, 2));
```

Or with curl:

```bash
# 1. Upload the file (or pass a public URL as "input" and skip this step)
curl -s https://platform.reducto.ai/upload \
  -H "Authorization: Bearer $REDUCTO_API_KEY" \
  -F file=@deck.pptx
# => {"file_id": "reducto://..."}

# 2. Run parse
curl -s https://platform.reducto.ai/parse \
  -H "Authorization: Bearer $REDUCTO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "input": "reducto://FILE_ID",
  "settings": {"model": "r-1"},
  "formatting": {"table_output_format": "md"}
}'
```

## Step 2: Read the result

On the sample (a synthetic 3-slide board deck), the job-specific part of the response:

```markdown
# Segment results ($M)

| Segment   | Q3 2025   | Q3 2026   | Change   |
|-|-|-|-|
| Apparel   | 182.4     | 196.0     | 7.5%     |
| Equipment | 141.9     | 155.3     | 9.4%     |
| Footwear  | 57.3      | 60.7      | 5.9%     |
| Total     | 381.6     | 412.0     | 8.0%     |
```

Slide titles become headings and slide tables come back as Markdown tables, ready for an LLM prompt.

[Full response format ->](https://docs.reducto.ai/parse/response-format)

## When to use something else

- **Parse**: to put decks and sheets into LLM context or a vector store.
- **Extract**: when you need specific values from a deck or sheet.
- **Reducto Office**: when an agent should edit the deck or sheet, not just read it.

## Where this goes next

- [PDF to Markdown for LLMs](/tools/pdf-to-markdown)
- [RAG pipeline for messy PDFs](/cookbooks/reducto-turn-a-complex-10-k-into-grounded-rag-with-r-1)
- [Reducto Office](/office)
- [Parse docs](https://docs.reducto.ai/parse)

Try one file free, no signup, with the [PowerPoint to Markdown](/tools/pptx-to-markdown) or [Excel to Markdown](/tools/xlsx-to-markdown) tool.
