> This page is for Document AI.

> Sarvam Voice Agents (Samvaad) is a voice AI agent platform for Indian languages: build, test and deploy AI calling agents on your own phone numbers, from ₹3 a minute. Start here: https://docs.sarvam.ai/voice-agents/why-voice-agents.
> For clean Markdown of any page, append `.md` to the page URL.
> For a complete documentation index, see https://docs.sarvam.ai/llms.txt. For a section index, append `/llms.txt` to any section URL (for example https://docs.sarvam.ai/voice-agents/llms.txt).
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.sarvam.ai/_mcp/server.

# Split a document bundle

> Break one big file into the separate documents inside it, labelled by type, and adjust the boundaries by hand.

**Split** finds where each document starts and ends inside one big file, then labels each part by type. A typical example is a loan application scanned as one PDF that holds a bank statement, a PAN card, and a GST registration certificate.

## Step 1: Add your file

Click **Split** under **Products** in the sidebar. Drag and drop your files, click **Upload documents**, or pick the **Loan bundle.pdf** demo document.

> **Note**
>
> Accepted formats: **PDF, JPEG, PNG, TIFF, ZIP**. Each run takes up to **10 files or 50 pages in total**. For larger bundles, use the [Split API](/api/api-guides-tutorials/document-intelligence/how-to/split).

## Step 2: Define document types

Under **Define document types**, add one row for each kind of document you expect in the bundle, with a **Document type** name and a **Description**. The setup works the same way as [Classify](/doc-ai/how-to/classify-documents): **+ Add type**, drag to reorder, **Form** or **JSON**, and **Load saved list**.

### Identifier keys

Sometimes a bundle holds two documents of the same type back to back, such as two bank statements from different accounts. To keep them apart, open the `⋯` menu on that type's row and click **Add identifier key**. Name the value that tells two documents of that type apart, for example the account number.

## Step 3: Run and review

Click **Run split**. The credits needed are shown under the button. When the run is **Done**, the **Output** tab shows how many pages and documents it found (for example *6 pages · 3 documents*). Each document appears as a group with its type, page range, and page thumbnails.

![Split output for Loan bundle.pdf reading 6 pages · 3 documents, with groups for Bank statement (pp. 1–3, 3 pages) and Identity document (p. 4), each with a type dropdown, an optional Name field, and page thumbnails.](/_fern-files/_img/422bd3cb428ade0065074c15300359374e3a990c5d6418e34fda5b0f7dbe1d0f.webp)

### Fix the boundaries

The result is fully editable, and changes save automatically:

* **Split a document.** Click between two pages to start a new document there.
* **Move a page.** Drag a page thumbnail into another document.
* **Leave a page out.** Click **×** on a thumbnail to drop that page, for example a blank separator sheet.
* **Relabel.** Change a document's type from its dropdown, and add an optional **Name**.
* **Reorder or delete.** Use the arrows to move a group up or down, and the trash icon to remove it.

Switch to **JSON** to see the raw output.

## Step 4: Download or reuse

* **Download** saves the results.
* **Save config** saves your document types for reuse. See [Save and reuse Configs](/doc-ai/how-to/save-and-reuse-configs).
* **View API code** gives you cURL, Python, or JavaScript for the [Split API](/api/api-guides-tutorials/document-intelligence/how-to/split).

> **Tip**
>
> Split pairs well with Extract. Split a bundle first, then run each part through the matching [Extract](/doc-ai/how-to/extract-fields) template, such as **Bank statement**, **PAN card**, or **GST return**.

Next: [Translate a document](/doc-ai/how-to/translate-a-document).