How to customize

Continue offers a handful of parameters in config.json that can be tuned to find the perfect balance between suggestion quality and system performance for your specific needs and hardware capabilities:

config.json

 "tabAutocompleteOptions": {
   "maxPromptTokens": 400,
   "prefixPercentage": 0.5,
   "multilineCompletions": "always"
 }

• maxPromptTokens: Sets the maximum number of tokens for the prompt, balancing context and speed.
• prefixPercentage: Defines the proportion of the prompt dedicated to the code before the cursor.
• multilineCompletions: Controls whether suggestions can span multiple lines ("always", "never", or "auto").

For a comprehensive guide on all configuration options and their impacts, see the Autocomplete deep dive.