Tutorials

Worked tutorials, run on your own machine.

Every tutorial assumes a working local setup. Start with setup if you haven't got a stack running yet.

Engineering

Adding a new model to an example package

Change the model in Marigold's chat example -- a newer open-weight model, a multimodal one, and a gated one requiring a HuggingFace token.

Engineering

Run a Local LLM Server with Docker Compose

Install the marigold CLI, start a local LLM server with Docker Compose, cache open-weight models and run a first application on your own hardware.

Engineering

Cursor Local Model: Connect a Self-Hosted LLM

Use a self-hosted Marigold model in Cursor through its OpenAI base URL override: the tunnel it needs, which features use it, and a VS Code route that stays on your machine.

Engineering

Local Document Search with Open WebUI

Ask questions of your own documents with every step on your own hardware: Marigold serves the chat and embedding models, Open WebUI handles retrieval.