How to Build a Private ChatGPT Using Open-Source Technology? Download our free white paper.

The fastest and most accurate
way to power your NLP projects

Natural Language Comprehension

  • High accuracy API

  • Sentiment and emotion analyses

  • Ready to scale

Natural Language Structuration

  • No-code ontology edition

  • Graph database

  • Turn information into knowledge

Auto Lettria

  • Data annotation with active learning

  • Train and deploy in minutes

  • No dev-ops required

Tell us more
about your Data

Our platform adapts to pretty much every text-related input, wether it is an email, PDF document or a Google Review. It nativaly embeds Optical Caracter Recognition (OCR) and parses tables.


Highlight and fine-tune
in a matter of seconds

Get started right away by creating a specific pipeline and choosing your analyses from among a baker’s dozen (i.e., spell check, sentiment analysis, IA entity recognition and more) Teach Lettria what you want to extract.


You’re not alone

Create a work group (or several) and invite your collaborators, so you can build dictionaries and annotate data together.

We also have a little gift for you: 750,000 pre-categorized words to help you kickstart your project. You’re welcome.


Start processing
& monitor quality

That’s pretty much it. You’re ready to go.
Our language processing toolkit is available as a single easy-to-use API and a no-code platform. Collect your results and monitor.

Read about our technological approach ->


Import your data, highlight what matters

Lettria allows you to upload your documents and tag the key pieces of information you want to extract.

Our active learning models will then learn from your examples to accurately extract targeted data from all your documents.


Edit your very own knowledge ontology
in a few clicks

Ontology is a fancy way to refer to a set of concepts related to each other.

In other words, just tell us what you precisely want to extract in your documents.


Generate properties & extract information in your document

Each property should be linked to at least one pattern. Each pattern is a sequence of elements to be found in the document.


Your data is now fully structured & accessible

You can export it in your own database, or in excel mode if you want. Lettria also offers graph database capabilities to help to explore and visualize your data in an innovative format.

Read about our technological approach ->


Tell us more about what you want to do

You can choose between binary and multi label classification, regression or word tagging.


Upload your raw data and unlock smart labelling

You’re not alone : invite teammates or clients!
Collaborate with peer-to-peer reviewing.
Accelerate with Active Learning (30% less annotation on average). Don’t have any labelled data?
That’s ok, some of our models work without.


We handle the
model training

Our system trains several models in parallel, depending on your needs and data, and selects the most efficient one.


Deploy in a matter
of clicks

Export your own custom model, or wait for us to send you an API token. No devops, no hassle.


Short on time? Ask our support team to help you kickstart your project.
Talk to an expert ->

Have a look at our platform stories


Find out how Sales teams use Lettria to extract key information from call transcripts and integrate into their CRM to improve business performance.

Logo La PosteLogo Julia
Vocabulary Enrichment

Find out how Lettria helps e-merchants improve search results on their website, deliver more relevant product recommendations, and increase their sales.

Logo Leroy Merlin
Web Data Analysis

Discover how marketing teams use Lettria to automate the analysis and synthesis of customer and employee surveys to identify key trends and information.

Logo VocazaLogo Uside
Build your NLP pipeline for free
Get started ->