Wispr Flow, Superwhisper, and Foil start from the same user desire: speak naturally and get usable text back. The useful comparison is not which tool is universally better. It is which tradeoff matches the way you work.

The short version

  • Wispr Flow is the best fit when you want a polished, cross-platform dictation and AI writing layer.
  • Superwhisper is the best fit when you want a mature local-first tool with a power-user mode system.
  • Foil is the best fit when you want Mac-first dictation with explicit provider choice, local transcription as an option, optional cleanup, and visible recovery when paste does not land.

Comparison table

Category Wispr Flow Superwhisper Foil
Best fit Cross-platform AI dictation and rewriting Local-first dictation for power users Mac-first dictation with provider control
Local transcription Cloud-first voice transcription Local models are central to the product Local whisper.cpp via localhost is supported
Hosted provider choice Wispr-managed service path Product-managed cloud features and local models Local, Groq, OpenAI, or custom OpenAI-compatible endpoints
Cleanup/rewrite Integrated AI writing layer Modes and workflows can transform dictation Optional cleanup with raw transcript fallback
Failure recovery Product-specific history and retry flows History and local recordings History, copy, paste, edit, export, retry, and clipboard fallback
Privacy posture Privacy Mode and retention controls, still server-processed for transcription Local processing with local recording/history behavior to understand Route-explicit: local when local, hosted when hosted
iOS expectations Public iOS app Public iOS app Closed iPhone preview; custom keyboard, Full Access, and build-scoped host-app proof only
Best reason to choose it You want polished cross-platform voice writing You want a powerful local dictation environment You want explicit Mac provider routing and recoverability

Why people compare Wispr Flow and Superwhisper

Wispr Flow made AI dictation feel mainstream: hold a key, speak naturally, and get cleaner text back. Its advantage is polish and breadth. The tradeoff is that voice transcription depends on Wispr's service path and internet connectivity. Wispr's own help center says Flow requires internet connectivity for voice transcription, and its support docs cover VPN, proxy, security-tool, and connection-loss cases.

Superwhisper appeals to a different user. It is local-first, powerful, and designed for people who want control. The tradeoff is that power can become configuration overhead. Users evaluating it should also understand its history and recording behavior, because Superwhisper docs describe recordings and metadata stored locally by default.

Foil sits between those poles. It is not a full cross-platform writing layer, and it is not trying to maximize modes. It is a Mac dictation app built around a simple pipeline: capture audio, choose a transcription route, optionally clean up the text, paste into the active app, and preserve the result if delivery does not work.

Local, hosted, or custom: the provider question

For many users, the real search is not "which dictation app is coolest?" It is "where does my voice go?"

Wispr Flow is cloud-first. That can be excellent when the service path is fast, available, and allowed by your network.

Superwhisper is local-first. That can be excellent when you want on-device processing and are comfortable with its configuration model.

Foil makes the provider route visible. You can use local whisper.cpp through a localhost OpenAI-compatible server, hosted Groq Whisper, OpenAI Whisper, or a custom OpenAI-compatible endpoint. That makes Foil useful for people who want a local baseline but still want hosted speed or their own infrastructure when it makes sense.

The important wording is "route." Local transcription, hosted transcription, and hosted cleanup are different privacy and reliability choices. A good Mac dictation setup should say which one it is using.

Paste recovery matters more than comparison tables admit

Most comparison pages focus on accuracy and price. Those matter, but a dictation tool can transcribe perfectly and still feel broken if the text does not land in the target app.

That is why paste recovery is a core Foil theme. Foil separates the stages: recording, transcription, optional cleanup, insertion, clipboard fallback, and history. If cleanup fails after transcription succeeds, the raw transcript should still be available. If paste does not land, the thought should not disappear.

No Mac dictation app can honestly promise perfect insertion into every target. The better product promise is narrower and more useful: make the delivery state visible and keep the transcript recoverable.

What about iOS?

iOS is where dictation promises get tricky. Keyboard extensions, secure fields, host-app behavior, app switching, permissions, background capture, and state reset can all affect whether text lands where the user expects.

Wispr Flow and Superwhisper both have public iOS apps. Foil's iOS work is a closed iPhone preview today. It uses a custom keyboard, requires Allow Full Access, and should only be described through build-scoped host-app proof. Build 12 currently has physical onboarding proof, Safari normal-field insertion proof, and Safari secure-field rejection proof. Notes and Messages need sterile surfaces before the build 12 rows can be rerun; Messages remains draft-only, Mail is deferred, and secure fields should reject the custom keyboard.

That evidence can become a product advantage. "Verified here, limited there" is more trustworthy than vague "works everywhere" copy.

Which one should you choose?

Choose Wispr Flow if you want a polished cross-platform AI dictation layer and you are comfortable with a cloud-first transcription path.

Choose Superwhisper if you want a mature local-first dictation tool and you are comfortable with deeper configuration, modes, and local history behavior.

Choose Foil if you mainly dictate on a Mac, want provider choice to be explicit, want local transcription available, and care about recovery when paste or cleanup fails.

For many people, the right answer is not permanent. Try the app that matches your current constraint: polish, local power, or route control.

Sources and further reading