LLM Usage

LLM helper (opt-in)

Opt in

• LLM calls are off until you configure a provider.
• Skip every prompt to keep running fully offline.

Providers and config

• repeater llm --set walks you through provider, base URL (when needed), API key, and model selection.
• Config is stored in a local auth file under your OS data directory (for example, ~/Library/Application Support/repeater/auth.json on macOS).
• OPENAI_API_KEY overrides the stored key for temporary runs.
• repeater llm --test confirms the config by listing models, repeater llm --clear forgets it instantly.

Ollama (local LLM) compatibility

• Ollama works because repeater speaks OpenAI-compatible APIs.
• Use repeater llm --set, choose “Other”, then set the base URL to http://localhost:11434/v1/.
• Leave the API key blank unless your proxy requires one.
• Make sure the model is available in Ollama (for example, ollama pull llama3.1), then select it from the model list.

Cloze generation

• Run repeater drill <deck>; if any C: cards lack [], repeater sends that text to your configured provider and patches the file before the drill continues.
• Leave the API key blank (or skip configuring a provider) to keep the feature idle.

Example cloze auto-fix:

C: The capital of Japan is Tokyo.

When the drill starts, the LLM will turn it into something like:

C: The capital of Japan is [Tokyo].

Question rephrasing

• Run repeater drill <deck> --rephrase to rephrase basic Q: questions before the session starts.
• The original answers are provided as context but are not revealed in the rewritten questions.

Example rephrase:

Q: What is the powerhouse of the cell?
A: The mitochondrion.

The LLM will rewrite the question only, for example:

Q: Which organelle is known as the cell's powerhouse?
A: The mitochondrion.