AutoKeras

Windows · Mac · Linux · Self-hosted

Freedom report

Three barsScore 6.5

Free tierA free tier is on its own pricing page
Open codeNo open-source code on record
Runs widely3 of 6 device platforms
DocumentedPlans, terms and facts published

AutoKeras is ranked #4 of 28 in AutoML software on Freedom251. It runs on Linux, macOS, Self-hosted, Windows. There is a free plan.

AutoKeras plans and pricing

All plans

AutoKeras Free Python package installed with pip autokeras.com · 4 Oct 2026

Compared on AutoML software

Feature engineering: Yes
Automated model selection: Yes
Workflow interface: code
Hosting model: self_hosted

Best AutoKeras alternatives

See all 12

Two bars FEDOT Free tier No. 1Free6.6 Two bars LightAutoML Free tier No. 2Free6.6 Two bars AutoGluon Free tier No. 3Free6.5 No signal Amazon SageMaker Autopilot Paid plans No. 5—6.4 One bar Auto-PyTorch Free tier No. 6Free6.4 One bar BigML Free tier No. 7$833.33/mo6.4

The record

Purpose: AutoKeras is an AutoML system based on Keras, developed to make machine learning accessible to everyone.
Tasks: Its supported tasks include image and text classification and regression, plus structured data classification and regression.
Multimodal: The documentation includes a tutorial for working with multi-task or multi-modal datasets.
Model design: Users can build custom models from AutoKeras blocks while letting the system search for detailed configurations.
Image features: Image workflows can configure architecture type, normalization, and data augmentation, or leave choices to be tuned automatically.
Model export: AutoKeras can export the best model found as a Keras Model.
Installation: The project provides pip installation instructions for the Python package.
Requirements: The installation guide says AutoKeras supports Python 3 and recommends PyTorch version 2.3.0 or later with the PyTorch backend.
Training behavior: By default, the last 20% of training data is used for validation, and users can provide a different split or their own validation data.
Resume runs: With overwrite set to false, running the same code again continues the previous fit from the directory contents.
Model selection: AutoKeras uses validation loss by default to select the best model, and supports customized metrics and objectives.
Support: The project directs users to GitHub Discussions to ask questions.
Maintainer: The package is developed by DATA Lab at Texas A&M University in collaboration with keras-team for version 1.0 and above.
Security: The official pages opened for this research do not state security certifications or compliance claims.
Price from: Free

Where it ranks on Freedom251

Best AutoML Software in 2026#4 of 28

Is AutoKeras yours?

Claim it for free: prove the domain, then correct facts, plans and screenshots. An editor reviews every change.

Claim it free Promote AutoKeras

Sources

autokeras.com· checked 4 Oct 2026
autokeras.com/tutorial/overview/· checked 4 Oct 2026
autokeras.com/tutorial/image_classification/· checked 4 Oct 2026
autokeras.com/tutorial/export/· checked 4 Oct 2026
autokeras.com/install/· checked 4 Oct 2026
autokeras.com/tutorial/faq/· checked 4 Oct 2026
autokeras.com/about/· checked 4 Oct 2026