dFL (Data Fusion Labeler) is a signal-centric, desktop and self-hosted platform designed for harmonizing, labeling, analyzing, and modeling complex, multi-sensor data sets.

Get Started Demos Visit Sophelio

About dFL

The Labeler, or dFL, is an all-in-one toolkit for data harmonization, labeling, and fusion. It quickly tames messy, multimodal streams—from sensors, simulations, or experiments—so any team can tag, organize, and transform them into clean, ML-ready assets. By automating preprocessing, enforcing consistent metadata handling, and prividing for data provinence, dFL accelerates data-driven discovery and production AI across industries.

Quick Start

Get up and running with dFL in minutes! Check out our Getting Started guide to begin.

Self-Hosted

dFL runs entirely on your infrastructure -- no cloud uploads or third-party dependencies required.

System Requirements

Ensure you have sufficient RAM for your dataset size, and Python 3.11+ for customization. Runs on Windows, Mac, and Linux.

---

Documentation Topics

1. Getting Started → Data Ingestion

• Data Provider and Utility scripts.
• LLM Prompting for automating ingestion and customization.

2. Quickstart Guide → Examples

• dFL walkthrough.
• Demos & examples.

3. Graphing & Analysis

• Supported graphs.
• Integrating custom graphs by writing simple python scripts.
• Exporting datasets from graphs in one click.

4. Data Harmonization

• Data Trim -- If your data set is large, trim first in the GUI for visualization & preprocessing.
• Data Filling options to automatically remove NaNs and empty data entries.
• Resampling techniques supported in the GUI.
• Smoothing techniques supported in the GUI.
• Normalization techniques supported in the GUI.
• Order of Operations and its importance in data harmonization.
• Custom Filters/Transformations can be readily added to the dFL GUI.

5. Data Labeling

• Manual Labeling -- How to manually label, export, and customize.
• Autolabeling -- Native autolabelers and configuration options.
• Custom Autolabeling -- Integrating custom python autolabelers.

6. User Customization

• Project Layouts -- Setting up new layouts, customizing, saving, and exporting.
• Themes -- Various themes are available in dFL.
• Colorblind options -- Palettes for colorblind users and publication.

7. Data Export

• Export from graphs -- Export custom harmonized datasets from graphs.
• Bulk export -- Export using bulk export parameters.

8. Premium Plugins (Store)

   a. Archaieus Integration

- A plugin that allows the extraction of smooth (up to 4^th-order) derivatives from noisy signals.
- A built in zero crossings autolabeler for identifying (up to 4^th-order) zero crossings.

   b. Direct Model Integration

- Coming Soon!