The Self-Serve API is in private beta. To learn more and sign up, please visit imerit.net/api
Introduction to iMerit
iMerit provides data scientists and machine learning companies a fully-managed data labeling solution. AI/ML teams work with iMerit experts to define plans of action for projects rannging from limited experiments to highly complex production-ready models. iMerit identifies how the project will be staffed, what annotation tool(s) would best serve the project goals and the edge case issues that customers might run into with their current project setup. Once both parties have agreed on the path forward, iMerit staffs the project, executes annotations and QA, and loops required parties in when edge cases arise.
iMerit's comprehensive offering covers six broad areas of the data annotation lifecycle:
- Project consultation
- Project setup and tool configuration
- Integration with your custom annotation tool (if required)
- Staffing, annotation and quality assurance
- Edge case coordination and resolution
- Project analytics
How the iMerit Self-Serve API works
iMerit’s Self-Serve API is the fastest and most direct way for teams to rapidly spin up and configure their iMerit projects. Once project creation is complete, our solution architects and delivery leads schedule time with your teams to deeply understand the outcomes of your projects and how we’ll work together to ensure those outcomes.
Using the API gives you access to the following functionality:
- Creating your project details
- Linking your data to the project
- Creating class ontologies
- Assigning annotation types to each class
- Assigning users to your project
- Uploading additional detailed project instructions
- Retrieving project analytics
- Downloading completed data
Multiple users can collaborate on shared projects by joining the same organization, which owns the projects. A subset of these user accounts can be given the role of organization owner, which allows those people to granularly manage access to the organization's projects using sophisticated security and administrative features.
A project holds the datasets to annotate and the ontology defining the annotations to perform. When the annotation process is started, metrics and analytics are automatically collected. Once the project is completed, the annotations can be downloaded.
The datasets are collections of media objects to annotate. Datasets belong to an organization and can be attached to multiple projects so they can be annotated more than once without having to duplicate the media objects. The media objects owned by a dataset are uploaded to the iMerit platform or imported from cloud storage providers.
If you're an iMerit customer and have any questions, please contact your account representative for help. To sign up for access to the private beta, please visit imerit.net/api