Welcome to distil labs

Distil labs provides a platform for training task-specific small language models (SLMs) with just a prompt and a few dozen examples. Our platform handles the complex machine learning processes behind the scenes. This allows you to focus on your use-case instead of managing large datasets and infrastructure.

Getting started

Install the Distil CLI:

curl -fsSL https://cli-assets.distillabs.ai/install.sh | sh

Minimal example

# Log in (if you dont have an account, use `distil register`)
distil login

# create a model for your specific task
distil model create my-first-model
# Output: Model created with ID: <model-id>

# Upload the logs and prompt of your agent
distil model upload-traces <model-id> --data ./my-data-folder

# Train a Small Model to solve your task as well as an LLM can
distil model run-training <model-id>

# Use your trained model
distil model deploy local <model-id> # deploy the model locally
distil model invoke <model-id> # get a script to invoke the model

That’s it! Your trained model is ready for local deployment. You can also use our Claude Skill to train models directly from Claude Code or Claude.ai.

Have structured data? You can also upload that directly instead of using traces.

Next steps

Ready to build your own specialized models? Continue to our How to train your SLM guide or explore detailed tutorials.

How to train your SLM?

Available models and training config

CLI Reference

API Reference