> For the complete documentation index, see [llms.txt](https://gdplabs.gitbook.io/glchat/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://gdplabs.gitbook.io/glchat/organization-admin/how-to-guides/pipeline-and-preset/dpo-pipelines.md).

# DPO Pipelines

<figure><img src="/files/hCQsPTYeXaXNpMO6TtPt" alt=""><figcaption><p><em>Fig 1: Available DPO pipelines</em></p></figcaption></figure>

You can add **presets** to the available **DPO (Document Processing Orchestrator) pipeline**. The DPO pipeline will be linked to the pipeline used by the **Knowledge Base**.

Select the desired pipeline from the available options on the DPO Pipelines page.

{% hint style="info" %}
Available options may vary and could differ from the example shown below.
{% endhint %}

***

## Add Preset

After selecting a pipeline, click New Preset to add a new configuration.

<figure><img src="/files/gAJOG5sp0EyfHEGBePwY" alt=""><figcaption><p><em>Fig 2: Add a new preset</em></p></figcaption></figure>

Enter a **Preset ID** to name the preset, then fill in the remaining fields. Click **Save** to store your changes or **Cancel** to discard them.

{% hint style="info" %}
The fields you see depend on which pipeline you selected. Find your pipeline in [Pipelines and Their Config Fields](#pipelines-and-their-config-fields) below. Its page explains every field it shows you, what values are accepted, and what happens if you leave a field empty.
{% endhint %}

<figure><img src="/files/zrkgn1MB2Y1WiiJ6fim1" alt=""><figcaption><p><em>Fig 3: Configure the preset</em></p></figcaption></figure>

***

## View and Edit Preset

Click the **Edit** icon in the **DPO Preset List** page to view or modify a preset. After making changes, click **Save** to apply updates or **Cancel** to discard them.

<figure><img src="/files/6s6b4R5cX3pLpOEI1BZ6" alt=""><figcaption><p><em>Fig 4: Edit a preset</em></p></figcaption></figure>

<figure><img src="/files/cpUZ23iik01dijsmtWis" alt=""><figcaption><p><em>Fig 5: Update preset preview</em></p></figcaption></figure>

***

## Pipelines and Their Config Fields

| Pipeline                                  | Choose it when you want to                                                                                                | Config fields                                                                                                               |
| ----------------------------------------- | ------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- |
| **Standard**                              | General-purpose pipeline. Start here if unsure.                                                                           | [Reference](/glchat/organization-admin/how-to-guides/pipeline-and-preset/dpo-pipelines/standard-pipeline.md)                |
| **Recommended**                           | Same as Standard, but extracts PDFs via Azure AI Document Intelligence for the best quality on scans and complex layouts. | [Reference](/glchat/organization-admin/how-to-guides/pipeline-and-preset/dpo-pipelines/recommended-pipeline.md)             |
| **Self-Hosted OCR** (`onprem-vision-ocr`) | Processes documents on your own infrastructure, using self-hosted extraction and OCR services instead of cloud providers. | [Reference](/glchat/organization-admin/how-to-guides/pipeline-and-preset/dpo-pipelines/self-hosted-ocr-dpo-pipeline.md)     |
| **Schema Based**                          | Ingests CSV files against a schema you define.                                                                            | [Reference](/glchat/organization-admin/how-to-guides/pipeline-and-preset/dpo-pipelines/schema-based-pipeline.md)            |
| **Schema-Driven Ingestion**               | Ingests web pages, extracting fields you select per domain from a shared schema template.                                 | [Reference](/glchat/organization-admin/how-to-guides/pipeline-and-preset/dpo-pipelines/schema-driven-ingestion-pipeline.md) |
| **Fulltext**                              | Indexes PDFs for keyword search only, without generating embeddings.                                                      | [Reference](/glchat/organization-admin/how-to-guides/pipeline-and-preset/dpo-pipelines/fulltext-pipeline.md)                |
| **Multimodal**                            | Same as Standard, but with LLM-based video captioning and multimodal indexing for richer media search.                    | Not documented yet                                                                                                          |
| **ID Regulation**                         | Purpose-built for Indonesian regulatory documents.                                                                        | Not documented yet                                                                                                          |
| **Recommended Graph RAG**                 | Same as Recommended, plus indexing into a knowledge graph alongside the vector store.                                     | Not documented yet                                                                                                          |
| **Graph RAG**                             | Same as Recommended Graph RAG, but extracts PDFs locally instead of via Azure.                                            | Not documented yet                                                                                                          |

{% hint style="info" %}
The pipelines available to your organization may differ from this list.
{% endhint %}