--- url: https://lettuceai.app/docs/quickstart title: "Quick Start — LettuceAI" description: "Set up LettuceAI in a few minutes. A guided first-run flow helps you connect a provider, add a key, pick a model, and start chatting." --- Menu # Quick Start The first time you open LettuceAI, a guided setup walks you through everything you need. This page mirrors that flow so you know what to expect. New to AI apps? If terms like tokens, context length, provider, or model are unfamiliar, read [AI Basics](/docs/ai-basics) first. The setup itself also includes a short, plain-language explainer for first-time users. ## Step 1: Welcome ![LettuceAI welcome screen](https://lhdgeo5fms.ufs.sh/f/m0TBUtMLsaiEGSaQBWolvZNz7b9S6JOnaWIErcYV8jTRAoKU) When you first launch LettuceAI, you will see the welcome screen. Tap **Get Started** to begin the guided setup. The welcome screen also has an **Other options** menu if you are not starting fresh: - **Sync from another device**: copy your chats, characters, and settings from a device that already has LettuceAI set up. See [Sync](/docs/sync). - **Restore from Backup**: load a previously saved `.lettuce` backup file. - **Skip for now**: jump straight into the app. Note that the app cannot generate replies until you have added at least one model, so most people should complete the setup first. Beta Build LettuceAI is currently in beta. If you encounter issues, please report them on our GitHub repository. ## Step 2: Tell it your experience level Setup first asks a quick question: have you set up an app like this before? - **No, this is new to me**: you get a short, plain-language explainer (about two minutes). It uses a simple car analogy. The app is the car, a **provider** supplies the engine, your **API key** is the key that starts it, and the **model** is the engine itself. It also explains that some models are free with a daily limit and others charge a small amount per message. - **Yes, I know how this works**: skip the explainer and go straight to provider setup, where you choose any provider, paste a key, test the connection, and continue. ## Step 3: Choose free or paid ![Choosing a setup path](https://lhdgeo5fms.ufs.sh/f/m0TBUtMLsaiERJD3yz4Jx2P34yDg6iKGIpAlcoTf9etwqHkz) First-time users are offered two easy starting points. You only ever pay the provider, never LettuceAI, and you can switch later without losing anything. - **Free (Google Gemini)**: costs nothing and never asks for a card, with a generous daily allowance. You can only use Google's own models, and mature or explicit content is blocked. - **Paid (OpenRouter)**: pick from almost every AI model through one account, with no daily caps. You pay only for what you use, and you need a card or prepaid balance to start. Not sure? Start free. You can always add a paid provider later from **Settings**. See the [API Keys guide](/docs/api-keys) for where to get a key from each provider. ## Step 4: Add your API key LettuceAI does not require an account. Instead you connect your own provider using an API key. Keys are encrypted and stored only on your device, never on our servers. Whichever path you picked, the setup walks you through creating a key with step-by-step screenshots, then asks you to paste it in: 1. Tap the button to open the provider site (Google AI Studio for Gemini, or OpenRouter for the paid path) and follow the guided screens to create and copy a key. 2. Paste the key into the box. 3. Tap **Connect**. LettuceAI checks that the key works before continuing. Copy the whole key Most "key did not work" errors are caused by a partially copied key. Copy it immediately when the provider shows it, as some providers only display the full key once. ## Step 5: Pick a model ![Model selection screen](https://lhdgeo5fms.ufs.sh/f/m0TBUtMLsaiEStjtW5fjxYOkrVMn2NmPXpy46uELTt0boAJw) A model is the specific AI that writes the replies. How you choose one depends on your path: - **Free (Gemini)**: a recommended model is set up for you automatically once your key is verified. - **Paid (OpenRouter)**: you are shown a short list of strong starter models with their current live prices. Pick one (the recommended option is a good default) and it is set up instantly. - **Experienced path**: choose your provider, then enter the exact model identifier, or pick from the provider's list. A model is required LettuceAI needs at least one working model before you can create characters or start chatting. You can add more models or switch between them at any time from Settings. ## Step 6: Choose your memory style ![Memory system selection](https://lhdgeo5fms.ufs.sh/f/m0TBUtMLsaiEgoDCLuz1uLhxGKCwTQRqZ4bmt2ikFOrvslNd) LettuceAI includes a memory system so characters can remember details across long conversations without resending the entire chat history every time. Pick how you want it to work: - **Dynamic Memory (Recommended)**: uses a small local embedding model to automatically retrieve the most relevant context. This keeps quality high in long chats and reduces token costs, with zero configuration. - **Manual Memory**: you pin and edit what gets remembered yourself, for total control. If you choose Dynamic Memory, the final step offers to download the embedding model (around 120 MB). It runs fully offline on your device. You can tap **Download & Enable** to set it up now, or **Skip for now** and enable it later in Settings. Recommended Dynamic Memory gives the best balance of quality, privacy, and lower cost. Learn more in the [Memory](/docs/memory) guide. ## Step 7: Start chatting ![Chat screen](https://lhdgeo5fms.ufs.sh/f/m0TBUtMLsaiEzapwjd5CgsDQbf3mrGI2U5jA9RHS8i0x1naJ) That is it. Setup is complete and everything runs locally on your device. Now you just need someone to talk to. 1. Create a [Character](/docs/characters) and a [Persona](/docs/personas), or import one and browse community characters from [Discovery](/docs/discovery). 2. Open the character and start the conversation. You're ready! LettuceAI is now fully set up. From here you can explore voices, image generation, group chats, and more. ## Where to go next Once you are past setup, a few places are the main hubs you will return to from anywhere in the app: - **Library**: a single view for everything you create locally. Filter tabs cover characters, personas, lorebooks, and images. New characters, personas, and lorebooks are created from here, including the AI-guided [Smart Creator](/docs/smart-creator). - **Search**: a fast finder over your characters and personas. Type to filter by name or description, then tap a result to start a chat or open the editor. - **Settings**: add more providers and [models](/docs/models), change your [memory](/docs/memory) setup, and tune the app. You can re-run pieces of the setup from here at any time. Privacy First Your API keys and chat data never leave your device. LettuceAI does not run its own servers and cannot see your conversations. [ PreviousInstallation ](/docs/installation)[ NextAI Basics ](/docs/ai-basics)