Not connected
v2.0 — Now with auto-upload

Ship your
ML models.
Like a pro.

Upload any scikit-learn model, get a public inference API on AWS in seconds. No config. No headache.

3
Models deployed
100%
Free tier
<1s
Avg latency
modelhub — inference
$ curl -X POST /predict \
-d '{"model":"iris","input":{
 "sepal_length":5.1,
 "petal_length":1.4}}'
# loading from s3://modelhub-models/iris.pkl
Prediction
Iris Setosa — 94.2%
✓ 200 OK  23ms  EC2 us-east-1
$
Featured models
Click any card to open its inference playground
View all
How it works
From local model to live API in 4 steps
01
Train locally
Build any scikit-learn model in Colab or Kaggle. Save it with joblib.dump().
02
Upload here
Drag and drop your .pkl file. Fill in a short model card — name, description, tags, author.
03
Auto-deploy
Backend validates the model, runs a test prediction, and auto-uploads to your S3 bucket via boto3.
04
Share & run
A public model card is created instantly. Anyone can run inference from the browser or via the REST API.

Upload a model

Drop a scikit-learn .pkl file — we handle the rest.

Drop your .pkl file here

or click to browse

Supports scikit-learn models saved with joblib • Max 50MB
📄
model.pkl
0 KB
Model Card
Suggestions:
Suggestions:
Suggestions:
Tip: Check your training script for accuracy score. Examples: wine: 100% | cancer: 95.6% | diabetes: R²=0.44
Suggestions:
These become the input fields in the playground
Press Enter to add a tag
Popular tags:
Model deployed!

Your model is live on AWS S3 and ready for inference.
A public model card has been created.

All Models

Community-deployed models running on AWS

Inference Playground DEMO
Prediction