Click any card to read the full documentation
Predict future values from historical time series — sales, demand, KPIs, stock levels. Train, compare and deploy with one click.
Label records, detect fraud, churn or intent — across any tabular dataset. Auto-selects the best algorithm for your data.
Segment customers, documents or products into meaningful groups — without labelled data. Visualise clusters directly in the platform.
Personalise product, content or action suggestions. From collaborative filtering to deep-learning sequential models.
Extract entities, classify text, summarise documents, translate and answer questions — powered by transformer models.
Classify images, detect objects, segment scenes and estimate depth — from ResNet to SAM to YOLO, all without writing model code.
Index your documents once and chat, search or query them via API. Auto-builds knowledge graphs with entity extraction.
Generate text, images and audio. Plug in any LLM or image model — OpenAI, Anthropic, Stable Diffusion and more.
Every model trained on the platform comes with built-in reasoning and explainability. See why a prediction was made, which features drove it, and how confident the model is — no black boxes.
See which input features have the most impact on your model's predictions — globally across all samples and locally for individual records.
Shapley values break down each prediction into per-feature contributions. Understand not just what the model predicted, but exactly why.
For tree-based models, visualise the exact decision path taken for any record — step by step, threshold by threshold.
Ask 'what would need to change to get a different outcome?' — generate actionable counterfactuals for any prediction automatically.
Every prediction includes a calibrated confidence score so you know when to trust the model and when to flag a record for human review.
Automatically detect performance disparities across demographic groups and surface potential bias before a model goes to production.
Ask the AI agent. Chat directly with your SHAP plots and model results — get plain-language explanations and concrete suggestions for improving performance.
Every time you train a model, we automatically generate the right evaluation charts — no setup needed. After deployment, the same applies to live usage: prediction distributions, drift detection and performance over time, all built for you.
Ask the AI agent. Chat directly with your training plots and usage data — get plain-language explanations and concrete suggestions for improving model performance.
Watch how fast you can train and deploy AI models — no code, no setup.
Search for a command to run...