AIRaML — 50 free machine-learning tools

Deployed Platforms

Three full apps — click to predict

Complete, multi-model platforms you can use end to end. For the 51 single-purpose tools that sit alongside them, see the toolkit below.

Platform

ML Unified Platform

Models

One app, four models. Select Iris classifier, Titanic survival predictor, Diabetes risk model, or Insurance premium estimator from a sidebar — all served from a single schema-driven FastAPI backend with dynamic forms.

ModelMulti-Model

Features26

4 Datasets · 26 Features

PlatformFastAPISchema-DrivenClassificationRegression

Exploratory Analysis

EDA Explorer

∞

Datasets

Upload any CSV dataset and instantly explore it — shape, dtypes, missing value heatmap, per-column distributions (histograms for numeric, bar charts for categorical), descriptive statistics, outlier counts, and a full Pearson correlation heatmap. No code required.

ModelPandas · NumPy

Features0

Any CSV

EDAStatisticsCorrelationDistributionsData Profiling

Vision

ML Vision Platform

150

Seg Classes

Three vision tasks in one app: classify images across 1000 ImageNet categories (MobileNetV2 · ResNet50 · SqueezeNet · GoogLeNet), detect objects with TinyYOLOv3 (COCO 80 classes), and segment scenes pixel-by-pixel with SegFormer-B0 (ADE20K 150 classes). All models run as ONNX on a FastAPI microservice.

ModelSegFormer-B0 · YOLOv3 · MobileNetV2

Features3

ImageNet · COCO · ADE20K

VisionONNXSegmentationDetectionClassification

All three run on a free Hugging Face Space — if it has gone idle, the first load takes a few seconds to wake up.

Three full apps — click to predict

ML Unified Platform

EDA Explorer

ML Vision Platform

51 tools, each live and testable