Resources to build
Legal AI .
We are curating a set of open source datasets and AI models to enable easy building of Justice solutions. These are curated from a variety of platforms such as Huggingface, Github, Govt. of India platforms and other open source initiatives.
Foundational AI blocks
Curated list of AI building blocks
Rhetorical Roles
An AI model pre-trained on BUILDNyAI to process the vast amount of information present in Court Judgments.
Learn more
Indian Legal BERT
An important resource for Legal NLP / AI-Law in India trained on 35000 Indian case laws
Learn more
Automatic charge identification
An AI model that identifies charges from the IPC, from a given text description of charges and facts of a criminal case.
Read
Court Judgement Prediction and Explanation
An explainable AI model that predicts the outcome of a court judgement
ReadTools
Awesome tools to develop justice solutions
Speechbrain
An open-source library to make speech recognition, speech-to-text, and speaker identification very easy.
Learn more
Microsoft Azure AML
An ML tool wherein you can create a model or use a pre-built model from an open-source platform.
Read
Google ML Kit
A mobile SDK wherein one can tackle common issues in apps or design new UIs easily using their NLP APIs.
Read
Rasa Open Source
An open source AI software used to build conversational UIs. Also boasts of pre-built models written in TensorFlow.
Read
Named Entity Recognition
This NER module can identify various legal entities mentioned in a judgment.
Read
TensorFlow
An open-source AI software that helps you develop and train predictive ML models.
Learn more
Markovify
A python library used to generate Markov chains on texts to generate random sentences. Useful when building a writing assistance tool.
Learn more
Tesseract Open Source OCR engine
An Open-source OCR platform that extracts text from images and documents and outputs it into a new file.
Learn more
Fritz AI
An ML platform for iOS and Android developers. It comes with pre-trained ML models baked right in and can be used to deploy custom-trained models.
Learn more
Obviously AI
It uses a unique technique called “Edge-Sharp AutoML” in order to create and train custom ML models.
Learn more
Apache Mahout
A distributed framework that manages data processing and can be effectively used for data mining.
Learn more
Teachable Machine
An AI experiment by Google that makes it fast and easy to create ML models without any coding/expertise.
Learn more
Create ML
An ML framework that allows the creation and training of ML models within the Apple ecosystem.
Read
OpenNN
A free, open-source AI software that works as a neural network library and deals with ML solutions.
Learn more
Google Cloud AutoML
Create a personalized ML model in a matter of minutes, with the aid of AutoML.
Learn more