kapi context suggest

$ kapi suggest

Read the files this project declares and file what kapi can see in them as suggestions:

• a word the files write more than one way, such as e-mail and email: the form written most often, instead of the others
• in a project that binds no voice, the words a starter voice pack avoids, when the files use three or more of them
• the files addressing the reader as you, when most blocks with a personal pronoun do and none says we

Each suggestion is noticed by kapi and waits for you in "kapi context review". Checks report it and none fails on it until you keep it. Nothing else is written: not the recipe, not a rules file. Without --ai no AI model is used.

--ai then asks the connected AI model about the same blocks, through term-extract and voice-infer, and files what it proposes the same way: the terms the files use, and the voice measures it reads, such as whom the files address and how long their sentences run. A term is filed only when the files write it as the model wrote it in two or more blocks. In a project that binds a voice, no voice measures are filed. These suggestions are noticed by kapi with the model's name. Without --ai the command needs no network; with it, kapi reaches the model's provider. With no AI provider configured, or no key for the one configured, --ai refuses before anything is filed; "kapi models setup" connects one.

Running it again files nothing new. A suggestion you dropped stays dropped, and one the files have changed is filed again as it now stands. --dry-run prints what it would file and records nothing; with --ai it still asks the model.

Commandcontext suggest

Groupcontext-person

Try it

Runs in your browser against a small sample file. Edit the command before running, or press Run to execute it as shown.

$ kapi context suggest

Flags

Flag	Type	Default	Description
--ai	bool	false	then ask the connected AI model and file what it proposes
--dry-run	bool	false	print what would be filed and record nothing
--project , -p	string		path to a kapi.yaml project recipe or its directory (auto-discovered from cwd if omitted)

More examples

$ kapi context suggest --dry-run

$ kapi context suggest --ai

$ kapi context suggest --json

← Back to the Command Reference