Guide
The terminal
The same agent, in a terminal, with the same account, sessions and settings as the app.
Starting it
| Command | Does |
|---|---|
letronna | open a conversation |
letronna -p "question" | ask once, print the answer and exit |
letronna --continue | pick up the most recent session |
letronna --resume | choose from the recent sessions; with an id, open that one |
letronna gui | open the window instead |
letronna setup | name, sign in, provider, model, install |
letronna --uninstall | take it off again |
--plan starts in plan mode: read and search only, never write or run. --yes runs tools without asking first — use it only where that is safe.
Slash commands
Type / in a conversation to see these:
| Command | Does |
|---|---|
/help | what Letronna can do here |
/new | start a fresh session |
/plan | read and search only, never write or run |
/model | show or change the model; add image, speech or video for those |
/gateways | see, add or remove gateways |
/team | bill hosted models to a team or to yourself |
/login | sign in to Letronna |
/logout | sign out on this machine |
/whoami | who is signed in |
/effort | how hard the model should think |
/thinking | show the model's thinking, or hide it |
/search | let the model search the web |
/jobs | commands still running in the background |
/skills | saved instructions you can hand the model |
/resume | pick up a session from before |
/where | show the project folder and session file |
/clear | clear the screen |
/exit | leave |