> ## Documentation Index > Fetch the complete documentation index at: https://docs.modulex.dev/llms.txt > Use this file to discover all available pages before exploring further. # Voice & dictation in chat > ModuleX chat is text-based. To speak a request instead of typing it, use your device's dictation to put text in the composer. Here is how, plus what the chat composer supports today. ModuleX chat works on text: you type a request, or **speak it using your device's dictation**, which puts the resulting text in the composer. This page shows how to use voice through dictation and what the chat composer supports today. ## What chat supports today The chat composer is the box at the bottom of every conversation. It does three things: Enter text and press Enter to send. Shift+Enter adds a newline. Messages can be up to a few thousand characters. Choose which language model answers — from ModuleX-managed models to your own provider keys. Send the message and watch the reply stream back live, including any tool steps the Assistant takes. ## Speak your request with device dictation To talk to ModuleX instead of typing, use your device's built-in dictation to enter text into the composer, then send it like any typed message. Open a chat and start your operating system's dictation — for example the microphone key on a phone keyboard, or your desktop's voice typing. Your device converts your speech to text directly in the composer, the same as if you had typed it. Press Enter (or the Send button). The reply streams back the same way it does for any text message. Device dictation is your operating system's feature. Your device converts speech to text locally or through its own service, and ModuleX only receives the resulting text — so accuracy, languages, and limits are entirely your device's. ## How chat answers your message However your text gets into the composer — typed or dictated — a single chat can answer it in three ways. You do not pick a mode; the chat decides based on what you ask and what you have connected. The default. An agentic chat that reasons, calls your connected tools step by step, and pauses for your approval when needed. Trigger a saved workflow from a message and watch the run unfold inside the thread. Ask a question and get an answer drawn from your connected knowledge bases, with the sources it used. Who can use chat: the chat surface requires an **owner** or **admin** role in the organization. See [roles and permissions](/security/roles-permissions). ## Where to go next How chat works end to end — the three ways a conversation is answered, and how replies stream. Pick which language model answers your chats, from managed models to your own keys. Attach a knowledge base and get answers grounded in your own documents. The agentic chat behind the default experience — how it reasons, calls tools, and asks for approval.