Text annotation tool

Text annotation tool

Kili is the text annotation tool, to help build training datasets faster, and solve NLP machine learning challenges that will impact your organization: chatbot utterances, emails, medical reports, voice of your customers, complex patents native pdf documents, raw txt. Kili annotation platform covers all major labelling tasks.

pdf text annotation labeling classification named entity recognition

You are in good company – small and large

Fast text annotation on all use cases

From simple conversations classification to advanced multi tasks interfaces with named entities recognition and relations extraction, all natural language processing use cases are covered. Each text annotation interface is designed to support productivity. Composability of each text labeling interface allows to adapt to the specific needs of each project.

For all tasks: text classification single class, multiclass, dropdrown for long list, nested classification to manage complex ontologies. Named entity recognition, relation extraction, Optical Caracter Recognition (OCR).

File types: txt or PDF.

Master labeled text data quality

Kili online or on-premise text labeling tool eases collaboration with domain experts or external labelers, and data-science teams. Labeled data quality monitoring is build-in thanks to advanced tools: consensus analysis, honeypot, review, and last but not least intructions.

Simple collaboration between labelers & data-scientists

Kili text annotation tool online for text, are designed to ease collaboration: on-board domain expert or external labeling teams, leverage data governance tools, advanced asset queue management distribution when it comes to labeling a large volume of data.

labeling collaboration text data

Work on a robust text annotation platform

Our platform is available online or on premise. For the community, we developed recipes, made open source on our Github. It makes it easy to leverage our powerful GraphQL API to interact programatically with the tool. It also allows to leverage the open source ecosystem: AutoSKlearn (open source autoML framework of SickitLearn) to train online a text classification model; Snorkel to implement weakly supervised learning strategies: for name entities recognition and classification, it allows to combine simple rules like dictionnaries or REGEX to build powerful pre annotators. Our wrapper in Python makes it even easier if your are not familiar with GraphQL.

Ready to simplify labelling in your company?

Discover the solution now