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

# Data Sources

> Connect your data to Jabrod pipelines

# Data Sources

A pipeline is only as good as the data inside it. Jabrod supports multiple data source types that you can add to your pipelines.

## Supported Types

| Type       | Description                                                                                    |
| ---------- | ---------------------------------------------------------------------------------------------- |
| **Text**   | Raw text pasted directly into the dashboard. Good for quick testing or small snippets.         |
| **File**   | Upload documents (PDF, MD, TXT, JSON, CSV). Jabrod automatically parses and extracts the text. |
| **URL**    | Provide a web URL. Jabrod will fetch the page and extract the textual content.                 |
| **API**    | Provide an API endpoint. Jabrod will fetch the raw response.                                   |
| **GitHub** | *(Prototype)* Sync readmes from GitHub repositories.                                           |

## The Ingestion Process

When you add a data source to a pipeline, it enters the ingestion queue:

1. **Status: Pending** - The source is registered. (For files, this means it was uploaded to cloud storage).
2. **Status: Processing** - Jabrod is currently downloading, parsing, chunking, and embedding the content.
3. **Status: Indexed** - The process is complete. The chunks are now searchable via the query API.
4. **Status: Error** - Something went wrong (e.g., unreadable PDF, broken URL).

If you change a pipeline's settings (like the chunking strategy), you can trigger a **Re-index** to reprocess all data sources using the new configuration.