Marco Hillger built the dictation app TypeWhisper after a stroke three years ago left him unable to type for sustained periods, he said in an interview. He had been using Wispr Flow, the market leader, and found that its lag made sustained work difficult. He wanted an input method that did not depend on a single model or service, so he started writing his own.

Hillger is 41, lives south of Berlin, studied communication and information management, and worked for six years as a developer in finance, where he said a two-person team grew to about 40 people before he started his own company. He built TypeWhisper first for his own daily work, and the commercial side came later.

Hillger said commercial licensing produced about EUR 2,000 a month after two months, while the core stayed free and open-source under GPLv3. The business page lists team pricing at EUR 19 a month for up to 10 devices and enterprise pricing at EUR 99 a month for support or a non-GPL commercial deployment. He described that as paid support around an open tool rather than access to a hosted service.

Key Takeaways

AI-generated summary, reviewed by an editor. More on our AI guidelines.

"I'm not an influencer"

Asked for a setup guide, Hillger declined to give one. He said he is not an influencer and does not produce that kind of recommendation. Pressed on his own preferences, he pointed to NVIDIA's Parakeet for local transcription and Groq for cloud transcription, and said he had switched the model he uses to build the app from Anthropic's Claude to OpenAI's Codex because it ran more consistently from day to day.

That hands-off stance is also the app's main usability problem. TypeWhisper exposes the speech engine, the cleanup model, per-app profiles, a dictionary, snippets, workflows and history, and a new user has to choose among them on first launch. The official TypeWhisper page lists WhisperKit, Parakeet TDT v3 and Apple SpeechAnalyzer as local recognition paths on newer Macs, with optional add-ons for Qwen3 ASR, Groq Whisper and OpenAI Whisper. Hillger said the app still needs clearer guidance, and that macOS is the stable version while Windows is in beta and iOS is still early.

The competition

Hillger's starting point had been a complaint about a rival. The latency in Wispr Flow, the best-funded app in the category, was what pushed him to write his own, he said. Wispr raised USD 25 million from Notable Capital in November 2025, TechCrunch reported, on reported total funding of USD 81 million and a valuation around USD 700 million per market-profile coverage. The company describes its goal as a voice operating system, and its security documentation notes that its Context Awareness feature is on by default on the desktop while its Privacy Mode and zero-data-retention controls are not.

He put Superwhisper closer to his own app. It runs local Whisper models, Parakeet and cloud providers behind custom modes, at roughly USD 8.49 a month or USD 249.99 for a lifetime license. Hillger called it the more polished of the two and TypeWhisper the more open.

In a follow-up note, Hillger said the platform owners are the structural threat he watches most closely. Apple now has SpeechAnalyzer, Copilot+ PCs have Fluid Dictation, and Gboard has voice typing built in. None of them, he wrote, yet ships the kind of configurable desktop workflow he is building.

He gave two examples of the work he means: a journalist who needs interview audio to stay on the machine until a quote is checked, and a developer whose cleanup model keeps rewriting code identifiers. He started TypeWhisper for that work after the stroke made it his own problem, and he was still shipping updates most days when we spoke.

The 2025 Buyer’s Guide to AI Dictation Apps (Windows, macOS, iOS, Android, Linux)
Voice typing finally rivals keyboards if you choose the right balance of privacy, speed, and cost.
The Implicator
Deepgram Launches Flux Multilingual Speech Model With 10-Language Mid-Call Switching
Deepgram today announced Flux Multilingual, a conversational speech recognition model that supports 10 languages with real-time language detection and the ability to switch languages during an active
The Implicator
Anthropic Adds Voice Mode to Claude Code, Starting With 5% of Users
Anthropic began rolling out voice mode for Claude Code on Tuesday, letting developers speak commands directly into the company's AI coding assistant instead of typing them. The feature uses a push-to-
The Implicator

FAQ

Frequently Asked Questions

What is TypeWhisper?

TypeWhisper is a local-first AI dictation app that lets users choose recognition engines, workflows, dictionaries, profiles and optional cloud cleanup.

Who created TypeWhisper?

TypeWhisper was created by Marco Hillger, a German developer who built it after a stroke made sustained typing difficult.

What makes TypeWhisper different from Wispr Flow?

Wispr Flow emphasizes polished cloud-first defaults and context awareness. TypeWhisper emphasizes local processing, engine choice, open-source licensing and user control.

How much does TypeWhisper cost for businesses?

Team pricing is EUR 19 per month for up to 10 devices, and Enterprise is EUR 99 per month for support or non-GPL commercial deployment.

Why does TypeWhisper matter if operating systems add dictation?

Basic dictation is moving into operating systems. TypeWhisper focuses on harder cases: privacy, specialist vocabulary, prompt control, workflows and app-specific behavior.

AI-generated summary, reviewed by an editor. More on our AI guidelines.

Analysis

San Francisco

Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: [email protected]