Voxtral Memos vs MicType

Record a voice memo. Get an instant transcription and a smart summary. It's that simple. Voxtral Memos uses Mistral AI's Voxtral speech model to turn your recordings into accurate, readable text — and then summarizes, extracts action items, or transforms them however you like. TRANSCRIBE ANYTHING Record meetings, lectures, ideas, or quick thoughts. Voxtral Memos transcribes them instantly using Mistral AI's state-of-the-art Voxtral model. Supports 13+ languages including English, German, French, Spanish, and more. SMART SUMMARIES & CUSTOM PROMPTS Go beyond transcription. Get concise summaries, extract to-do lists, create journal entries, or define your own custom prompt templates. Every memo can be transformed in multiple ways. BUILT ON EUROPEAN AI Powered by Mistral AI, Europe's leading AI company based in Paris. Your data is processed by a GDPR-compliant, European technology provider. PRIVACY BY DESIGN - Powered by Mistral AI — a European, GDPR-compliant AI provider - No account required — ever - No tracking, no analytics, no telemetry - Zero third-party SDKs - All data stored locally on your device - Your API key is secured in the iOS Keychain - Audio is only sent when you explicitly request transcription - Fully open source on GitHub BRING YOUR OWN KEY Voxtral Memos uses a Bring Your Own Key model. You connect your own Mistral API key, so you're always in control of your usage and costs. No subscriptions, no hidden fees. FREE & OPEN SOURCE The app is completely free. No ads, no paywalls, no subscriptions. The full source code is available on GitHub under the MIT license. If you enjoy the app, you can leave a tip to support development. REQUIREMENTS - A Mistral AI API key (free to create at console.mistral.ai) - iOS 26.0 or later

MicType turns your own AI into a keyboard. It runs on your account, not ours: you need an OpenAI API key and a connection. Without those, MicType cannot transcribe. Speak, and it types Tap the mic and talk — in any app. Your words are recognized as you speak, and clean, punctuated, ready-to-send text lands at the cursor. Hold the mic instead and say what you want done — polish, translate, rewrite, draft a reply, answer a question — and the result is typed right where you are. Your AI, your key, no middleman Connect OpenAI with your own API key. Your audio and your text go straight from your phone to OpenAI. No account, no sign-in, and no MicType server — we never receive, relay or store any of it. Before anything is sent, a plain-language disclosure tells you exactly what leaves your phone; remove your key and nothing is sent. Accurate first Every sentence is recognized twice — live while you speak, and once more from the whole recording when you stop — and the more accurate result is used. Text usually lands two to four seconds after you stop. Polish that sounds like you Fillers and stutters go, misheard words get fixed, numbers are written the way you write them, and a long rambling note becomes clear points. Facts, questions and negations are checked before the text is inserted — if a rewrite drifts, you get your own words instead. Teach it your words Add names, products and jargon to your vocabulary and they are spelled right from the start, in the recognizer itself. MicType also notices the words it keeps getting wrong and suggests them. Buy once, no subscription Free for 30 days, then a single one-time purchase — no monthly, no yearly. Separately, you pay your own provider for the tokens you use. Optimized for Chinese and mixed Chinese-English, and works in 20+ languages including Arabic. For people who want AI-assisted typing without handing their data to one more company.