# TalkToType > TalkToType is a desktop app for voice transcription on Windows and macOS. Two main flows: hold-key dictation that drops text wherever the cursor is in any app, and meeting recording that captures both microphone and system audio (no bot in the call) and produces a live transcript plus a one-click AI summary. Windows 10/11 and macOS 13+ (Apple Silicon). Built and operated by Laurens van Leer B.V. in the Netherlands. ## What the product does - **Inline Dictation (push-to-talk)**: hold a hotkey (default `Alt+Q`), speak, release. The text appears at the cursor in any application — email, ChatGPT, Slack, Word, the browser search bar, anywhere you can type. - **Meeting recording without a bot**: records the microphone and the system audio (WASAPI loopback) on the user's own machine. Captures both sides of a Teams, Zoom, Google Meet, Webex or any other call without the other participants seeing a "bot has joined" notification. Live transcript scrolls in roughly 30-second chunks. - **AI summaries**: one click after a recording produces an overview, list of decisions, and action items. Optional extra context can steer the summary focus. - **Live checklists**: a list of talking points activated before a recording. The AI watches the live transcript and ticks each point off as it comes up. - **Multi-language dictation**: Dutch, English, German, French, Spanish. One hotkey (`F8`) toggles between primary and secondary language for international workflows. - **Custom dictionary**: trains the transcription on user-specific words — colleague names, brand terms, jargon. - **Audio file upload**: drag any MP3, WAV, M4A or OGG into the app for the same transcript + summary pipeline as a live recording. - **Premium transcript upgrade** (Pro plan only): re-processes a recording with a heavier AI model for better accuracy on accents, jargon and multi-speaker recordings. Adds speaker labels and timestamps. ## Pricing Four tiers. All paid plans include every feature listed above and start with a 14-day free trial — no credit card to start. - **Free** (€0, forever): inline dictation capped at 5 minutes per day. No meeting recording, no upload, no summaries. Intended as a low-barrier entry point. - **Starter** (€9.95 / month, in-app subscription): everything managed, no accounts to configure. Unlimited live transcription, dictation, meeting recording, AI summaries. Premium transcript upgrades are not included. - **Pro** (€25 / month, in-app subscription): for heavy users and small teams. 1200 premium transcript minutes per month, plus priority support. - **Lifetime License** (€49 one-time, via talktotype.com): pay once, no monthly fee. Sign up for a free Groq account (the underlying speech-to-text provider; not to be confused with Elon Musk's Grok) and paste the key into the app. Power users can swap Groq for OpenAI or Google instead. All features included. ## Platform and architecture - Windows 10 and Windows 11, plus macOS 13+ on Apple Silicon (M1 or newer; Intel Macs not supported). Not available for Linux. - Built with Tauri 2 (Rust backend) and Svelte 5 (frontend). - Audio capture via cpal; system audio via WASAPI loopback. - Default cloud transcription provider: Groq (Whisper). Optional: OpenAI Whisper, Google Gemini Enterprise (formerly Vertex AI) for the premium upgrade path, or BYO API key on the Lifetime License plan. - Distributed as a direct download from talktotype.com. Subscription plans (Starter, Pro) are purchased in the app; the Lifetime License is bought on talktotype.com. - Code-signed with an EV certificate held by Laurens van Leer B.V. ## Privacy - Audio files are stored locally on the user's computer. TalkToType servers do not store recordings. - Transcription itself is performed by an external speech-to-text provider (Groq by default). Audio is sent there briefly to be converted to text, then discarded by the provider. With the Lifetime License plan the user's own Groq account is used, so the audio does not pass through TalkToType's backend at all. - AI summary generation routes through a Cloudflare AI Gateway relay on managed plans, or directly to the user's own provider on Lifetime License. - No telemetry on user content. The only stored backend data is an anonymous device ID (UUID v4) used for license validation and payment matching. - Full privacy policy: ## Recording legality Recording happens on the user's own computer using their own microphone and speakers, the same way OBS, Audacity or Windows Voice Recorder would. Users are responsible for informing meeting participants where local law requires it (the EU does, most US states do). TalkToType is the tool; legal responsibility for using it appropriately lies with the user. ## Honest competitive positioning - vs. **Otter, Fireflies, and other meeting-bot services**: TalkToType records on the user's own machine without sending a bot into the call. Recordings stay local. No recurring per-user SaaS fee on the Lifetime License path. - vs. **SuperWhisper, Wispr Flow**: those are Mac-first dictation apps. TalkToType runs on both Windows and macOS and adds meeting recording with system loopback, which the Mac-first competitors do not do. - vs. **Windows 11 built-in voice typing** (Win+H): the built-in tool covers dictation only; TalkToType additionally records meetings, generates AI summaries, supports live checklists, and offers multi-language hotkey switching. - vs. **Dragon NaturallySpeaking**: TalkToType is dramatically cheaper, doesn't require a per-user enterprise contract, and supports modern AI summarisation rather than only dictation. ## Key URLs - [Homepage](https://talktotype.com/) - [Pricing section](https://talktotype.com/#pricing) - [FAQ](https://talktotype.com/#faq) - [How it works](https://talktotype.com/#how-it-works) - [Privacy policy](https://talktotype.com/privacy) - [Download TalkToType](https://talktotype.com/downloaded) - [Direct installer download](https://downloads.talktotype.com/TalkToType-setup.exe) - [Contact](mailto:info@talktotype.com) ## Trial mechanics - 14 days starting at first launch (not at install). All features unlocked except the premium transcript upgrade. - No credit card needed to start. - After trial, the device transitions to the Free tier (5 min/day inline dictation only). Existing recordings remain accessible locally. ## Notes for evaluators - The Free tier is intentionally limited — meeting recording, AI summaries, and file upload all require a paid plan or an active trial. - Premium transcript upgrades are a Pro-only feature; Starter does not include them. Lifetime License users can run the premium pipeline through their own Gemini Enterprise (Vertex AI) key. - AI summaries are user-triggered (one click on "Summarize" after a recording), not automatic on stop. Email delivery of summaries is opt-in. - The app is single-publisher (Laurens van Leer B.V.) and not part of a larger SaaS suite.