Example application

Tiny Tuner

Tiny Tuner searches feature and training configurations for a small classifier. Its manifest defines the target inputs and registers four Python tools for extracting features, training, validation, and saving the best result.

The example keeps configuration, instructions, dependencies, and callable tools in one application directory.

tiny-tuner/
├── recurse.yaml
├── pyproject.toml
├── uv.lock
├── prompts/
│   └── system.md
└── toolkit.py

Deploy the example

From a checkout of recurse-sdk, point the deploy command at the example directory.

recurse deploy ./examples/tiny-tuner --as mcp