Understand, Curate, Expand

Build better models faster by easily understanding and improving your dataset. Explore puts advanced search, data exploration, data curation, and tools for dataset improvement at your fingertips.

Explore and Understand Your Data

Use similarity search and natural language search powered by vectors to navigate your dataset or filter through annotations, predictions, and metadata to find exactly what you need.

Efficiently Expand Your Dataset

One-shot Annotation

Select a single instance of a polygon or bounding box and leverage advanced algorithms to automatically find and suggest annotations for similar instances across your entire dataset.

Batch Classification

Use similarity and NLP search to classify large volumes of data quickly. Search for items, select them in bulk, and apply the desired class efficiently.

Act and Automate with Orchestrate

Define and configure custom actions in Python to apply to your data and run them directly from Explore. These actions can include writing annotations to storage, applying custom rules and QA, running models, or using LLMs for judgment setups. Enjoy total freedom to tailor actions to your specific needs.

Ensure Quality at Scale

Ensure quality by using search tools to review every instance of a class across your dataset. Use vector search to find and correct similar items with errors. Approve, disapprove, or reclassify items and send feedback to annotators or reviewers.

Vector Search

Quickly identify wrong annotation patterns.

Rating System

Approve or disapprove LLM actions and leave comments.

Assign Tasks

Simply forward tasks to specific team members or annotators.

Rating System

Approve or disapprove the LLM actions, leave comments.

Create Visibility and Collaboration

Create and share subsets of items (images, text) with colleagues or partners using signed URLs. Recipients can review the items without needing a software license, facilitating easier collaboration and data sharing within your organization.