Text Classification API
Build your own custom text classifier with machine learning, using your own data. Completely no-code, and all in one place.

Spend less time juggling data and tools, and more time fine-tuning your model.
You already have your data, your annotation campaigns, your team and your knowledge base on Lettria.
Now you can train a machine learning model easily, within your project where everything (and everyone) else is already there.


Test and evaluate multiple different models, and pick the best one.
You can launch multiple trainings, and have multiple models in one project — which you can then test and evaluate.
Compare your models, choose the best one, and iterate on it further. All in one place.
Easy machine learning with AutoLettria
You can create custom text classification models using your own data on the platform, no coding required.
• Train your own ML models
• Test, evaluate and iterate on multiple models
• Easily deploy your ML model to production
%20(1).png)
Patrick Duvaut
Head of Innovation
%20(1).png)
Patrick Duvaut
Head of Innovation
%20(1).png)
Patrick Duvaut
Head of Innovation
Frequently Asked Questions
Ideal for text classification scenarios like claims triage, sentiment analysis, ticket routing, and content moderation—particularly when handling proprietary or regulated data.
Absolutely. Once models are trained, you can deploy them easily via API, ensuring predictions are accessible in production workflows .
Yes. AutoLettria supports multiple concurrent training sessions, allowing you to test configurations and select the best-performing model within your project
No. Business users and domain experts can drive model creation: upload data, define labels, annotate samples, then trigger training and deployment—no code required
AutoLettria trains small, efficient BERT-based models that capture critical text features. These models can then fine-tune larger LLMs, offering high performance with reduced compute cost
By analyzing annotated documents, AutoLettria automatically generates models in minutes. Users can launch multiple training runs, evaluate performance, and iterate—all within the same project
AutoLettria is Lettria’s no-code AutoML tool that enables users to train custom text classification models using their own annotated data—without any coding or machine learning expertise