X/Twitter | LinkedIn | Docs | Website | API
At Tanaos we develop task-specific Small Language Models for various text processing tasks. The table below lists the available models along with links to several ways to use them.
For ease of use and maximum performance, we recommend using them through the Tanaos API. With 100ms end-to-end latency, our API is perfect for real-time applications. Get a free API key by signing up on our platform.
If you'd rather use the models on your own metal, or if you want to fine-tune them on your specific tasks without training data, use our open-source Artifex library.
Additionally, all models are available for download from our Hugging Face page.
| Model | Description | API | Artifex | HuggingFace |
|---|---|---|---|---|
| Text Classification | Performs general-purpose text classification based on the user requirements. | - | Artifex | - |
| Guardrail | Flags unsafe, harmful, or off-topic messages. | API | Artifex | HuggingFace |
| Intent Classification | Classifies user messages into predefined intent categories. | API | Artifex | HuggingFace |
| Reranker | Ranks a list of items or search results based on relevance to a query. | - | Artifex | - |
| Sentiment Analysis | Determines the sentiment (positive, negative, neutral) of a given text. | API | Artifex | HuggingFace |
| Emotion Detection | Identifies the emotion expressed in a given text. | API | Artifex | HuggingFace |
| Named Entity Recognition | Detects and classifies named entities in text. | API | Artifex | HuggingFace |
| Text Anonymization | Removes personally identifiable information (PII) from text. | API | Artifex | HuggingFace |
| Spam Detection | Identifies whether a message is spam or not. | API | Artifex | HuggingFace |
| Topic Classification | Classifies text into predefined topics. | API | Artifex | HuggingFace |