Navi Chip Functions

Navi Chip Functions

Navi Chip Functions — what a chip can call

A chip does not run off on its own — it runs inside Navi, and everything it does with her goes through one doorway: navi_internal, her own chip API. At the top of root.py the chip brings her in and takes hold of the live instance:

• import navi_internal — brings Navi's chip API into the chip.
• navi_instance = navi_internal.navi_instance — the live Navi herself. Every function below is called on this instance; take hold of it first, at the top of run() or at module level.

Talking and thinking. The three functions a chip reaches for most:

• print_message(text, include_ai_name=True) — prints a message to Navi's terminal, with her name in front of it unless you pass include_ai_name=False. This is the main way a chip talks: results, progress, errors — if the operator should see it, it goes through here.
• llm_chat(user_message, called_from_app=False) — sends a message to Navi's model and returns its response, so a chip can ask her to think and then work with what she says. The published navi-nmap chip uses it to turn raw scan output into a summary a person can read.
• process_message(user_message) — runs a message through Navi's full pipeline — command detection and all — exactly as if the operator had typed it. Reach for it when the chip wants Navi herself to handle something, not just her model.

The difference between the last two matters: llm_chat is a question and an answer; process_message is handing her the keyboard.

The operator, the terminal, and the session

Who she is working for:

• get_user() — returns the current operator's name: the name selected in Settings, the one her greetings use.
• set_user(sys_user) — sets the current operator's name.

The screen and the session:

• clear_terminal() — clears the terminal screen and reprints the Navi banner. A chip that is about to draw something full-screen starts here.
• chat_with_navi() — opens an interactive chat session with Navi, dropping the operator into conversation with her from inside the chip.
• setup_history() — initializes the history for prompt sessions, so a chip that runs its own prompts keeps them navigable.
• setup_navi_vocab() — registers commands and aliases with Navi's vocabulary (the spaCy entity ruler), so her command detection learns the words the chip brought with it.

This page is their home. The full method list lives here in the documentation, complete and kept current — Chip Studio carries no method panels of its own, so the list is written once and read while you write. Upstream, the project's chip-developer wiki page, Developing Chips — Indepth, documents navi_internal.py's NaviApp class — the source this page follows. New to chips entirely? Start with the Chips development guide, in the sidebar.