The Data – Parser node processes unstructured input and automatically routes it into structured output types. It serves as a multipurpose entry point for raw data of unknown or mixed types and splits them into defined formats such as text, tables, audio, images, or video.
Input
- Data – Accepts a wide range of file or content types including binary files, mixed media, or serialized formats. This is the only input for this node
Outputs
- Audio – Outputs audio content if present in the input (example .mp3, .wav)
- Image – Outputs any image content (example .jpg, .png, base64)
- Table – Outputs structured tabular data (example .csv, .xlsx, HTML tables)
- Text – Outputs plain text or extracted textual content from documents
- Video – Outputs video content, including video-only or video-with-audio files (example .mp4, .mov)
How to Use
- Connect a data source to the Data input (example files from File System, emails, documents)
- The parser will automatically detect content types embedded in the input
- Use the appropriate output ports (Audio, Image, Table, Text, Video) to route the parsed results downstream
- No manual configuration is required for format detection
Notes
- This node acts as an auto-routing gateway for pipelines that require pre-processing of heterogeneous inputs
- If the file contains multiple data types (example a PDF with embedded images and text), multiple outputs may be activated simultaneously
- It is typically used at the early stage of pipelines for ingestion and content breakdown
