Datature
No-code computer vision platform for labelling image data, training models and deploying them to cloud or edge.
Datature is a computer vision platform that covers the full path from image annotation to model training and deployment without requiring code. Datature is used by uploading visual datasets, labelling them with AI-assisted tools, configuring training workflows in a drag-and-drop interface, and serving the trained model through an API or on edge devices.
Best for Teams that want to build and ship custom computer vision models without writing their own training and deployment code.
- Listed in
- Image Recognition & OCR
- Platforms
- Web, API
- Pricing model
- Freemium
- Free plan
- Yes
- API
- Yes
- Made by
- Datature
- Added
Last verified: . Information taken from datature.io.
What is Datature?
Datature is a platform for building vision AI, organised around three products named Label, Train and Deploy. In the labelling stage, teams annotate images and video with bounding boxes, segmentation masks, keypoints and 3D annotations, helped by model-assisted tools and by review and consensus steps that keep labels consistent. In training, a workflow is assembled visually, hyperparameters such as batch size and epoch count are adjusted, and results are compared against ground truth. Supported tasks include classification, object detection, pose estimation with keypoints and pixel-level segmentation. Trained models can be exported in multiple formats or hosted on Datature's GPU infrastructure for real-time and batch inference, with edge deployment on higher tiers. The company also sells consulting services such as proof-of-concept projects. Four plans exist: Free, Developer, Professional and Enterprise, each with quotas for images, storage and compute minutes.
Key Features
Annotation Tools
Labels images and video with boxes, masks, keypoints and 3D annotations, with AI assistance to reduce manual work.
Consensus Workflows
Routes annotations through approval and agreement steps so several labellers produce consistent data.
Visual Training Workflows
Builds a training pipeline by dragging blocks for the dataset, augmentation and model, with no code.
Hyperparameter Tuning
Exposes settings such as model architecture, batch size and number of epochs for adjustment.
Model Evaluation
Shows predictions next to ground truth so performance can be inspected visually.
Hosted Deployment
Serves a trained model through an API on GPU infrastructure for real-time or batch inference.
Screenshots
Use Cases
- Defect Inspection
- Train a detector that flags faulty items on a production line.
- Medical Image Segmentation
- Annotate scans such as DICOM files and train segmentation models on them.
- Retail Shelf Monitoring
- Recognise and count products for inventory tracking.
- Crop Monitoring
- Analyse field imagery to assess plant condition and yield.
- Traffic Analysis
- Detect and track vehicles and people in city camera footage.
Things to know
- Quotas apply to every plan
- Each plan includes monthly compute minutes for training and inference, and more can be purchased if needed.
- Image size limit
- Images are limited to 100MB each unless you are on the Enterprise tier or use your own storage.
- Paid plans are quoted
- Developer, Professional and Enterprise plans show a contact button instead of a listed price.
- Academic programme
- The site has a separate pricing page for academics and researchers.
FAQs
What is Datature?
Datature is a no-code platform for annotating visual data, training computer vision models and deploying them.
Is Datature free?
Datature has a Free plan for individuals and hobbyists that includes up to 300 images, 300MB of storage and 300 minutes of training compute per month. Larger plans are arranged through the sales team.
What computer vision tasks does Datature support?
Datature supports classification, object detection, keypoint-based pose estimation and segmentation.
Does Datature have an API?
Yes. Datature says deployed models integrate with existing systems through its APIs, and it publishes developer documentation.
Can Datature models run on edge devices?
Datature lists edge device deployment alongside cloud deployment, with edge fleet management among its plan features.
How does Datature count video?
For video files, Datature counts each frame as one image credit, so a 120-frame clip uses 120 credits.
Who it's for
- Machine learning engineers
- Computer vision teams
- Researchers
- Healthcare imaging teams
- Manufacturers
Comments
No comments yet. If you have used Datature, tell other readers how it went.
Alternatives to Datature
| Tool | Free plan | Starting price | API |
|---|---|---|---|
UltralyticsComputer vision platform from the makers of YOLO for annotating data, training models on cloud GPUs and deploying them. |
Yes | — | Yes |
 CVATData annotation platform for labelling images, video, 3D point clouds and audio, offered as cloud, self-hosted and open source. |
Yes | — | Yes |
 Segments.aiData labeling platform for robotics and autonomous driving that annotates 3D point clouds and camera images together. |
— | — | Yes |
 Digitap.aiAPI platform for banks and fintechs covering digital KYC onboarding, bank statement analysis and alternate-data credit assessment. |
— | — | Yes |
 Dog Breed FinderFree AI tool that estimates a dog's breed from one uploaded photo and shows the three closest matches. |
Yes | — | — |
Listed in the same category as Datature. A starting price appears only where we checked it on the maker’s pricing page; a dash means not checked or not published. All Image Recognition & OCR tools