Guide
How to use Letronna
From making an account to the terminal's every command. Something missing? Write to support@letronna.ai.
Getting started
Using Letronna
- The appThe Letronna window: sessions, models, settings, and what differs on Windows, Linux and macOS.
- The parts of the windowWhat is where in the Letronna window: the top bar, the sidebar, the chat, the composer, the status bar and the panel.
- First startThe setup page the window shows when no model is connected yet.
- The side panelThe Files, Changes and Jobs tabs beside the chat.
- The command paletteOne box to run an action, open a page, find a session, pick a model or attach a folder.
- Keyboard shortcutsEvery key in the Letronna window, part by part.
- SettingsEvery page of the window's Settings, and what each option does.
- AppearanceThe window's themes, film grain and motion.
- VoiceSpeaking a message into the composer, and having replies read aloud.
- Quick EntryA small bar to ask Letronna something from anywhere, without opening the window.
- Tray, notifications and updatesThe tray icon and its menu, what happens when you close the window, desktop notifications, and checking for a new version.
- Asking first and modesWhen Letronna asks before it acts, the six modes, plan mode, and what is refused in every mode.
- What the agent can doThe tools Letronna uses: files, commands, web search, reading documents, attaching pictures and controlling apps.
- SessionsFinding, resuming, rewinding, branching, summarising and exporting conversations.
- MemoryWhat Letronna remembers between sessions, when it saves, where it keeps it, and how to see and change it.
- Skills and your own commandsSaved instructions Letronna can follow, and prompts you run as slash commands.
- Artifacts and previewsThe files Letronna makes for you, and previewing pages and pictures.
- Scheduled jobs and loopsPrompts that run on a timer, sessions that keep working toward a goal, and being told when a run ends.
- Helpers and workflowsLetronna handing parts of a task to helpers, running workflows of many helpers, and working in a git worktree.
- The terminalletronna in a terminal: how to start it, its options, and every slash command.
- Commands and optionsEverything letronna takes on the command line: its commands and its options.
- Every slash commandAll of the terminal's slash commands, grouped by what they do.
- Keys in the terminalTyping, editing, history, copying and stopping in the terminal.
- Scripts and one-shot runsletronna -p: asking once from a script, its options, JSON output and running with nothing set up.
- The screenWhat the terminal shows, from the top of the window to the footer under the box.
- Approving tool callsThe card that asks before Letronna acts in the terminal: the diff, the answers, and questions from the model.
- Lists and pickersHow lists, switches and sliders work in the terminal's box: picking a session, a model, a mode or an effort.
- Selecting and copyingMarking text with the mouse in the terminal, where the copy goes, and copying over ssh.
- ScrollingMoving back through the conversation in the terminal with the wheel and the keys.
- Pasting and picturesPasting long text into the terminal's box, and attaching pictures from the clipboard or by dragging them in.
- Title and notificationsHow the terminal's tab title and desktop notifications tell you Letronna is working, done, or waiting for you.
- Colours and textThe terminal's colours, what each means, turning colour off, and how wide text is laid out.
- Terminals that workWhich terminals Letronna's screen works in on Windows, macOS and Linux, and what differs between them.
- The mobile appLetronna on iPhone and Android: what it does, and where each part is explained.
- Signing inSigning in to the mobile app, and signing out of other phones and browsers.
- ChatsStarting chats, finding them again, and what holding a message or a chat does.
- ModelsChoosing the model a chat uses: hosted models, and models on the iPhone.
- Working offlineChatting with no internet on an iPhone: Apple Intelligence and models you download to the phone.
- Adding a modelAdding a model from Hugging Face to run on your iPhone.
- Web search and photosHow answers search the web, and sending photos in a chat.
- Live Activity and notificationsFollowing an answer while you are in another app: the Dynamic Island, the lock screen and notifications.
- Your computerPairing the phone with Letronna on your computer, chatting with it, and managing its accounts.
- Code sessionsChanging a GitHub repo from the phone: each message runs in Letronna Cloud and opens a pull request.
- SettingsWhat each part of the mobile app's Settings does.
- GatewaysWhere Letronna gets its models: your Letronna account, OpenAI, OpenRouter, Ollama or any OpenAI-compatible server.
- Hosted models and creditLetronna's own models: how credit works, what Pro changes, and making images, speech and video.
- Local modelsRunning a model on your own computer: why you might, and what you need.
- OllamaUsing a model you run with Ollama: install, pull a model, pick it in Letronna.
- Other local serversAdding LM Studio, llama.cpp, vLLM or any OpenAI-compatible server as a gateway of your own.
- Choosing a modelWhich local model fits your memory, why it must call tools, and the context window.
- Working offlineWhat Letronna can and cannot do with no internet connection.
- Serving your accounts to other toolsThe Local endpoint: an OpenAI-compatible address on this computer that other tools can use your connected models through.
Connect and extend
- Telegram, Discord and SlackTalking to Letronna from a chat app on your phone, and getting scheduled jobs' reports there.
- MCP serversGiving Letronna more tools from MCP servers: adding them, signing in, and choosing which tools it may use.
- Plugins, hooks and moreProject instructions, plugins, hooks, language servers, and running Letronna inside Zed or JetBrains.
- WalletThe wallet the agent pays from: paying sites that ask to be paid, sending money, and the limits you set.
Cloud and teams
- Letronna CloudRun a task on Letronna's servers, keep a service running, or host a web app with its own link, all from the chat.
- Tasks in the cloudSend a task to Letronna Cloud from the chat or the terminal, follow it, carry it on and stop it.
- ServicesKeep a bot, a scheduler or a webhook receiver running in Letronna Cloud, day and night.
- Web apps with a linkA web app Letronna builds in the cloud stays up at its own address, wakes on a visit and sleeps when nobody comes.
- Looking after themSee what your services and web apps print, and take them down or bring them back, from the chat or the terminal.
- A terminal in the browserOpen Letronna itself, running in Letronna Cloud, in a terminal on the Cloud page.
- Telegram while your computer is offHave Letronna Cloud answer your Telegram chat when this computer is off.
- What Letronna Cloud costsHow tasks, services and web apps are charged, and where to see what you spent.
- LimitsHow much Letronna Cloud runs at once, how big and how long, and what it does not allow.
- TeamsOne credit pool for 2 to 5 people: seats, spending limits, the audit log, and who pays.
Account and data
Developers
Help