Instant Dataset Insights
See rows, columns, data types, summary statistics, value counts, and a correlation heatmap the moment your file finishes uploading.
No Code · No Account · 68 Models
Turn Raw Spreadsheets Into Trained ML Models in Minutes
Upload a CSV or Excel file, explore and clean it, then train and compare up to 68 machine-learning models side by side. Export the full report when you're done.
- No account needed
- Data expires after 24 hours
- Export PDF, data, and models
Compare Models · iris.csv · Classification
| # | Model | Accuracy | F1 Score | Time |
|---|---|---|---|---|
| 1 | Support Vector Classifier (RBF) | 96.67% | ||
| 2 | Gradient Boosting Classifier | 96.67% | ||
| 3 | Logistic Regression | 93.33% | ||
| 4 | K-Nearest Neighbors Classifier | 93.33% | ||
| 5 | XGBoost Classifier | 93.33% |
Best Model
Support Vector Classifier
96.67%
accuracy on 30 held-out rows
- Rows
- 150
- CV Mean
- 96.67%
- Built-In Models
- 68
- Learning Tasks
- 3
- Export Formats
- 4
- Lines of Code Needed
- 0
Features
Everything You Need to Go From Data to Decisions
A complete machine-learning workflow in your browser: insights, cleaning, training, comparison, and export, all in one place.
Smart Data Cleaning
Find missing values column by column, preview exactly what a fix will change, and apply it in one step. Your original data is always kept.
68 Models, Recommended for You
Pick classification, regression, or clustering and get a shortlist tuned to your dataset's size, from Logistic Regression to XGBoost.
Honest Training Progress
Watch each model train with real, measured timings. Stop after the current model at any time, and see exactly why a model failed.
Side-by-Side Comparison
Rank models on any metric in a leaderboard, bar chart, or radar chart, with the best score in every column highlighted for you.
One-Click Export
Download a PDF report, the cleaned dataset, session metadata, and every trained model as a ready-to-load .pkl file.
How It Works
From Upload to Insight in Four Steps
No setup, no installs, and no notebooks. Every step explains what it's doing and what to do next.
- 1
Upload Your Data
Drop in a CSV or Excel file up to 50 MB, or try a bundled sample dataset. No sign-up required.
- 2
Explore and Clean
Review statistics and missing values, then preview and apply cleaning until your data is ready to model.
- 3
Train and Compare
Choose a target column, accept the recommended models or pick your own, and compare every result side by side.
- 4
Export Everything
Download the full PDF report, the cleaned dataset, and trained models to use in your own projects.
Model Library
68 Models, Recommended for Your Data
Choose what you want to predict and MLInsights suggests the models most likely to work well for a dataset of your size.
30
Classification
Predict a category, such as churn or no churn, spam or not spam.
- Logistic Regression
- Random Forest
- XGBoost
- LightGBM
- CatBoost
- SVM
- KNN
- Naive Bayes
- MLP
- Stacking
28
Regression
Predict a number, such as a price, a demand forecast, or a score.
- Linear
- Ridge
- Lasso
- ElasticNet
- Random Forest
- Gradient Boosting
- XGBoost
- LightGBM
- SVR
- Huber
10
Clustering
Discover natural groups in data that has no labels at all.
- K-Means
- Mini-Batch K-Means
- DBSCAN
- Agglomerative
- Spectral
- BIRCH
- OPTICS
- Gaussian Mixture
Exports
Take Every Result With You
Everything you create in MLInsights can be downloaded and reused, from a polished report to production-ready model files.
PDF Report
Dataset overview, missing values, and every model's metrics in one shareable document.
Cleaned Dataset
.csv / .xlsx
Your data after cleaning, in the same format you uploaded, ready for other tools.
Session Metadata
.json
Settings, features, targets, and results for every model, in a machine-readable file.
Trained Models
.zip of .pkl
Every fitted model, including its scaler, plus a manifest describing the expected inputs.
Privacy by Design
Your Data Stays Yours
MLInsights collects only what it needs to run your analysis, and keeps it only as long as you need it.
No Account, No Tracking
There's nothing to sign up for, and the site runs no analytics or advertising trackers.
Expires After 24 Hours
Uploaded files are kept only so you can keep working. After 24 hours they expire and are permanently erased.
Delete Anytime
Remove one dataset or all of them instantly from Settings, without waiting for expiry.
Never Sold or Shared
Your data is used only to run the analysis you ask for. It's never sold or used to train anything else.
Read the full Privacy Policy to see exactly what's stored, where, and for how long.
FAQ
Frequently Asked Questions
Quick answers about how MLInsights works, what it supports, and how your data is handled.
Do I need to write any code?
No. Everything happens in a visual interface: upload a spreadsheet, review insights, clean the data, and train models with a few clicks. No Python or notebooks required.
Do I need to create an account?
No. Each upload creates a private session linked to your browser. You can manage or delete your sessions at any time from Settings.
What file formats and sizes are supported?
CSV, XLSX, and XLS files up to 50 MB. Comma, semicolon, tab, and pipe-separated CSV files are all detected automatically.
How long is my data kept?
Your data expires 24 hours after upload. Expired data can't be opened again and is permanently erased by automatic cleanup. You can delete it sooner at any time from Settings.
How does MLInsights choose which models to recommend?
Recommendations depend on the task you pick and the size of your dataset. Small datasets get models that work well with less data; large datasets get models that scale, such as LightGBM and XGBoost.
Can I use the trained models outside MLInsights?
Yes. Export the models ZIP to get each fitted model as a .pkl file you can load with joblib in Python, plus a manifest listing the columns each model expects.
Are the results reliable enough for real decisions?
Metrics are measured on a held-out test split and checked with cross-validation, but they're estimates. Treat them as a strong starting point and validate important decisions with a domain expert.
Ready to See What Your Data Can Do?
Upload a dataset and get insights, trained models, and a full report in minutes. No account and no setup.