A voice input app developed by SuperUltra, Inc. in Canada. It runs on macOS, Windows, and iOS, executing Whisper-family speech recognition models on the device itself so that more than 100 languages can be transcribed even offline. It is not tied to a particular editor: press a shortcut key, speak, and the text is inserted wherever your cursor happens to be ─ Slack, a browser, Cursor, or Claude Code. Rather than writing out exactly what you said, it can pass the result through an AI step called Super Mode to shape the text ─ “make it read like an email,” “turn it into bullet points” ─ in the same pass. Since its release in July 2023, it has expanded into translation, transcription of audio files, and meeting recording.
Key Features
- Voice input that works in any app: It runs in the background and is invoked with a keyboard shortcut. Anywhere you can type, you can dictate ─ chat, documents, code editors, and terminals all behave the same way
- On-device processing and offline operation: Because the speech recognition model runs on your machine, transcription works even without a network connection. Audio can stay on the device, which suits work notes and confidential material
- Support for 100+ languages: Japanese and many other languages are supported. You can write out what you said in the original language, or translate any language into English as you dictate
- Modes that shape the output: Alongside built-in modes such as plain transcription and formatted text, you can build your own custom modes. By changing the instructions passed to the AI, you can prepare outputs for email, meeting notes, bullet lists, and other uses
- A choice of cloud and local AI models: You can switch between cloud models from providers such as OpenAI, Anthropic, and Groq and local models that run on the device, and you can bring your own API keys
- Audio/video file transcription and meeting recording: Beyond live dictation, existing recordings and meeting audio can be turned into text in bulk
- Custom vocabulary (dictionary): Proper nouns and in-house terminology that tend to be misrecognized can be registered in advance to improve accuracy
- Security and compliance: The service states compliance with SOC 2 Type II and HIPAA, reflecting a design intended for business use as well
Pricing
| Plan | Price | What it includes |
|---|---|---|
| Free | $0 | Basic voice input features with no time limit. Pro features can be tried for up to 3,000 words |
| Pro (monthly) | $8.49 / month | Bring your own API keys, unlimited use of cloud and large local models, translation into English, audio/video file transcription, priority support |
| Pro (yearly) | $84.99 / year (equivalent to 2 months free) | Same features as the monthly plan |
| Lifetime | One-time purchase (see the official site for the price) | All Pro features plus future updates, purchased outright |
| Enterprise | Contact sales | SOC 2 Type II, centralized billing, model controls |
A student discount and a 30-day refund guarantee for paid plans are available. A single license covers both the desktop and iOS versions.
Pricing is current as of August 2026. Please check the official site for the latest information.
Pros & Cons
✅ Pros
- Because it works regardless of the app, it fits into daily use as a general input method rather than something reserved for a single task
- On-device processing means it keeps working offline, including while traveling or on an unstable connection
- The AI shapes the text to fit the context instead of only transcribing, which reduces the amount of manual cleanup afterwards
- The free plan covers the basics indefinitely and lets you sample Pro features up to a limit, making it easy to evaluate
- A lifetime option exists, so the longer you use it the lighter the cost is compared with a subscription
⚠️ Cons
- Serious use of cloud models and larger local models requires the Pro plan; the free tier is limited in scope
- Since local models run on your machine, quality and speed depend on device specifications (memory, CPU/GPU)
- In fields with strong speaking habits or heavy jargon, some upfront work is needed to register vocabulary and tune modes
- Android and Linux are not supported ─ only macOS, Windows, and iOS
- AI shaping is convenient, but you still cannot skip checking that the meaning has not drifted from what you intended
Comparison with Similar Services
| Criteria | Superwhisper | Wispr Flow | macOS built-in dictation | OpenAI Whisper (the model) |
|---|---|---|---|---|
| Form | Background app (macOS / Windows / iOS) | Background app (macOS / Windows / iOS) | OS feature (Apple devices) | Open-source model |
| Offline operation | Yes (local models) | Mainly cloud processing | Processed on device | Possible if you build the environment yourself |
| AI text shaping | Yes, via Super Mode | Yes | No (inserted as spoken) | No (needs to be combined separately) |
| App coverage | Works in any app | Works in any app | Works in any app | Depends on implementation |
| Pricing | Free plan; Pro as subscription or lifetime | Free plan; paid tiers are subscription | Free | Free (compute costs are separate) |
The first question is whether the OS’s built-in dictation is enough. If you only need text to go in, the built-in feature will do; a dedicated app earns its place when you want formatting, dictionaries, purpose-specific modes, and file transcription as part of a daily input method.
Who Is It For
- Writers, planners, and support staff who produce a lot of text and want to reduce typing
- People who prefer to speak a rough draft in one go and then tidy it up
- Anyone turning meeting or interview recordings into text to shorten the work of writing minutes
- People who would rather not send audio externally, or who need to work offline
- Developers who frequently give long instructions to AI agents while coding
Summary
Superwhisper combines on-device speech recognition with AI text shaping, aiming to turn speaking into an everyday input method. Its practical strengths are the background design that responds the same way in any app and the fact that it keeps working offline. On the other hand, core features such as unlimited cloud model use and file transcription sit behind the Pro plan, so it is worth starting with the free plan and the 3,000-word trial to see whether it suits your speaking style and use cases, then deciding whether the monthly or lifetime option makes more sense.