Customize how context is delivered to your LLM. These settings affect chat and API responses.
The default model used for chat and API requests.
Maximum tokens in the context window (1024-128000).
Maximum number of documentation chunks to include in context (1-100).
Number of characters to overlap between chunks for better context (0-1000).
Controls randomness in responses (0.0-2.0). Lower values are more deterministic.
Optional custom system prompt to prepend to all requests. Leave empty to use default.