PixelDock

No API fees · Runs on your device · Trainable Naive Bayes

Train your text classifier

Teach a local Naive Bayes classifier using your own labeled examples.

Local models, learning & document helpers

Choose a local model, trainable classifier or statistical helper. Inputs stay on your device; model files download only when you run a model tool.

Text & settings

0 / 16,000 characters

Up to 64 KB · stays on your device

Analysis result

Enter your text or try a sample.

Method & limitations

This tool uses the named method. Its output includes the method and limitations.

Runs in a Web Worker. Inputs and results are cleared when you leave or choose Clear text; they are not saved in browser storage. Downloaded files remain on your device.

A guide to this tool

How to use Train your text classifier

Go to the tool ↑
  1. Enter your text or load this tool’s sample.
  2. Choose the available settings and fill in a second input if this tool has one.
  3. Run the analysis and review the output and method notes before using it.

A sample you can try

Sample input

Product photos for the online store

Training examples (JSON)

[
  {
    "label": "commerce",
    "text": "product photos and online shopping"
  },
  {
    "label": "commerce",
    "text": "store product prices and purchase"
  },
  {
    "label": "sport",
    "text": "football team and sports game"
  },
  {
    "label": "sport",
    "text": "team training and football match"
  }
]

Sample output

commerce — 96.0%
sport — 4.0%

Naive Bayes trained only on your examples in memory. Relative weights are not validated accuracy, calibrated confidence or a basis for consequential decisions.

Read the method, limits and tool-selection guide →