Skip to main content

Label data.
Train your model.

From your first labeled dataset to a model you can download. Build with Data Labeling and Train, which share projects, datasets, and a team balance.

Quickstarts

Get your first labeled dataset

Label 10 support tickets by intent with a rule-based pipeline and download the finished file.

Train your first model

Train a model on 10 support tickets, review the cost, and download its first Model Version.

Have the agent build your project

Describe the data you need in plain language and review the plan the agent drafts before you deploy.

Work from the CLI

Install the toloka CLI to sign in, set up projects, and run pipelines and Training from your terminal.

Explore the platform

Data Labeling

Build a pipeline, work with experts or an LLM, and export your results.

Train

Fine-tune with your examples or compress a prompt with gisting.

Label, train, and evaluate

Turn a labeling export into training data and score predictions against a golden set.

SDKs and API

Define pipelines in TypeScript, script the CLI, or call the HTTP API.

Admin

Manage team access, balance, and billing.

Troubleshooting

Diagnose a run by symptom and look up exact error messages.

Reference