Universal Data Tool
Search…
Universal Data Tool
Universal Data Tool Docs
Installation
Running On-Premise
Collaborative Labeling
Building and Labeling Datasets
Image Segmentation
Image Classification
Text Classification
Named Entity Recognition
Entity Relations / Part of Speech Tagging
Audio Transcription
Data Entry
Video Segmentation
Composite Interfaces
Landmark / Pose Annotation
Importing Data
Upload or Open Directories
Import File URLs
Import COCO Images
Import from Google Drive
Import from AWS S3 Bucket
Import from CSV or JSON
Import using AWS Cognito
Import Text Snippets
The Format .udt.json
What is the .udt.json format?
What is the .udt.csv format?
Machine Learning
Jupyter Notebook Integration
Import Datasets into Pandas
Fast.ai
Integrate with Any Web Page
Integrate with the Javascript Library
Getting Started with React
Train your Workforce
Getting Started with UDT Courses
Frequently Asked Questions
Powered By GitBook
Running On-Premise
Run the Universal Data Tool on your infrastructure.
For small companies or projects, you might want to consider just using the Desktop Application or udt.dev, which requires less setup!
The Universal Data Tool can be run and customized on your own infrastructure easily. This is best for users who want to keep their version of Universal Data Tool stable, and don't need new features from weekly releases. You may also want to keep your data within your network.

Run via Docker

The Universal Data Tool builds docker containers on every new release. You can run the Universal Data Tool using...
1
# Starts the Universal Data Tool on port 3000 and a collaborative server on 3001
2
docker run -d -p 3000:3000 -p 3001:3001 \
3
-e UDT_collaborationServer_url=http://localhost:3001 \
4
universaldatatool/universaldatatool
Copied!
You can provide different options to the docker container to control various aspects of Universal Data Tool. Below is a table of a few of the configuration parameters can customize. You'll know if a configuration value is used because it will appear in the starting logs like the following image...
Configuration values are shown in a table on start
Environment Var
Description
UDT_collaborationServer_url
Collaboration server URL
UDT_pluginUrls
Comma-delimited URLs to plugins that should be loaded by default
UDT_auth_proxy_corsproxy
Proxy to use for CORS requests, e.g. to make api requests that are disallowed by CORs
UDT_auth_s3iam_accessKeyId
Access Key for uploading/importing from AWS S3 buckets
UDT_auth_s3iam_secretAccessKey
Secret Access Key for uploading/importing from AWS S3 buckets
UDT_s3iam_region
Region of S3 Bucket
UDT_auth_cognito_identityPoolId
Cognito Identity Pool Id
UDT_auth_cognito_region
Cognito Region
UDT_auth_cognito_userPoolId
Cognito user pool id
UDT_auth_cognito_userPoolWebClientId
Cognito user pool web client id
UDT_auth_cognito_storage_awsS3_bucket
Bucket to use for Cognito object storage
UDT_auth_cognito_storage_awsS3_region
Region of bucket to use for Cognito
UDT_labelhelp_disabled
Disable remote paid collaboration requests
UDT_labelhelp_apikey
API Key for requesting work from Label Help
These are the most common configuration values, but you can also customize hotkeys or any other attribute the appears in the AppConfig.

Independent Collaboration Server

Sometimes, you may want to run the Collaboration Server on a different machine than the Client Application, you can do this by using the npm module.
1
# via npm
2
npm install -g udt-collaboration-server
3
​
4
udt-collaboration-server --port 3001
Copied!
Previous
Installation
Next
Collaborative Labeling
Last modified 1yr ago
Copy link
Contents
Run via Docker
Independent Collaboration Server