our technology

Natural Language Processing

More accessible than ever

Companies generate an immense amount of text, which then they need to process repeatedly through manual tasks (filling in databases, classifying documents, copy/pasting…) The cost in time and resources is high. However most of these tasks can be performed automatically with Natural Language Processing (NLP). This technology can assist you in:

  • automation of processes
  • analysis (decision-making)
  • communication (chatbots)

In Avantopy, our mission is to make NLP more accessible than it’s ever been with our custom, as-a-service solutions. No need to hire new talent, develop new technologies or assign in-house resources.

Our hallmarks are low risk experimentation and easy integration of customized solutions. We focus on two broad tasks (automatic text classification and information extraction) that are at the core of a variety of applications (process and document management, semantic search, CV matching…). We use state-of-the-art methods (machine learning, deep learning) for maximum accuracy and our in-house annotation software for fast and painless data preparation.

NLP is the subfield of Artificial Intelligence (AI) that deals with human language. Its goal is to train computers in understanding and using languages so that they can perform text-based tasks.​

We have one goal: we process unstructured free text and derive actionable information. We can solve a large variety of business problems (process and document management, semantic search, CV matching…) by focusing on two clearly defined technical solutions. 

Avantopy Tag

Classification API

Classification is the process of providing additional data to a document, a paragraph, a sentence, … this additional data is also called metadata. Avantopy Tag retrieves metadata about your documents and performs analyses such as:

  • Sentiment analysis
  • Topic classification
  • Genre classification

Avantopy Extract

Information Extraction API

Information extraction looks at what is said in a document and makes that information explicit. Avantopy Extract digs out content from your documents and performs analyses such as:
  • Information extraction
  • Named entity recognition
  • Monitoring mentions

APIs for easy integration and constant improvement

Secure access

Securely access your deployed machine learning models


Scale happily and robustly thanks to Kubernetes

Constant improvement

We integrate new technology when it comes available, if you upgrade, all you'll notice is that it works better than before


Our in-house annotation software delivers fast, minimal and painless data preparation.

  • built for teams
  • web-based
  • we provide training and initial setup
  • made to iterate (quality control, revision)
  • intelligent selection of examples​


We use state-of-the-art methods in machine learning for maximum accuracy.

  • No existing models? We build them from scratch.
  • We adapt existing models to work best for your type of documents.

Our technology stack