Superwhisper vs Dollop.

• Cross-platform. Mac, Windows, and iOS — same workflow across devices. Dollop is Mac-only and will stay that way.
• Older Macs. Superwhisper runs on most Macs since the Whisper model carries the speech recognition. Dollop requires Apple Silicon and macOS 26.
• Maturity. Superwhisper has been shipping for years. Larger user base, more workflows, more "I asked Reddit how to do X" answers documented online.
• Cloud option. For very long dictations Superwhisper can optionally route through cloud models for speed. Dollop does not — it's on-device only by design.
• Wider language support today. Superwhisper's Whisper-based stack covers more languages than Apple's speech recognition does in some niches.

• Free vs $249. No subscription, no one-time purchase, no credit card. Every feature unlocked from install.
• 0 GB RAM cost. Apple Foundation Models lives in the OS. Superwhisper loads ~1.5 GB of Whisper into your RAM whenever it runs.
• Tuned cleanup adapter. Superwhisper uses generic LLM polish. Dollop uses a LoRA trained specifically on dictation cleanup pairs — fewer hallucinations, better identifier handling, more conservative edits.
• Per-app tone matching. Dollop reads the active app and applies Casual / Formal / Excited tone presets. Superwhisper has one cleanup style for all apps.
• Architecturally simpler. No Whisper model to download, manage, or update. No cloud-mode toggle to remember to keep off. Apple ships the foundation model with the OS.
• No model-update friction. Superwhisper requires you to download a new Whisper model when better ones release. Dollop updates only the cleanup adapter (~150 MB) when Apple ships a new foundation model with macOS.

Both apps are "on-device." The architectural difference is who hosts the model. Superwhisper ships Whisper as part of the app — every Mac running it loads 700 MB to 3 GB into RAM (depending on which model preset you pick). Dollop uses Apple Foundation Models, which Apple keeps warm in the OS regardless of whether your dictation app is running. The net effect: Dollop's app contributes 0 GB of RAM beyond the OS baseline. On a 16 GB Mac with Cursor, Chrome, Slack, Docker, and Figma already open, that's the difference between "it works fine" and "the system is swapping."

Superwhisper's cleanup pass uses a general-purpose LLM. It works, but generic LLMs over-edit dictation — they smooth tone in ways the speaker didn't intend, hallucinate phrases that weren't said, or rewrite sentences that were already fine. Dollop's cleanup adapter is a LoRA fine-tuned on thousands of dictation pairs (raw transcript → cleaned text). It behaves more conservatively: drops fillers, adds punctuation, preserves identifiers — but doesn't restructure your sentences. For dictation specifically (rather than general writing), this is the right tradeoff.

Different cost structure. Superwhisper has to ship its own model, host downloads, support multiple platforms, and maintain a Whisper integration as Whisper itself evolves. Dollop uses Apple Foundation Models — Apple ships and maintains the model. Our marginal cost per user is essentially zero. We don't have a cloud bill, a model storage bill, or a per-platform engineering bill. So we don't need to charge.

Apple Silicon + macOS 26 is a hard requirement. Older Macs and Intel Macs cannot run Dollop — Superwhisper handles that case better. We're also Mac-only on purpose; the entire architecture depends on Apple Foundation Models, so there will never be a Windows or iOS version. Superwhisper is the right pick if cross-platform matters. And while Dollop's cleanup adapter is genuinely better-tuned than generic LLM polish, Superwhisper's larger Whisper models have an edge in less-supported languages. For English and major European languages, the experience is comparable; for niche languages, Superwhisper's wider model lineup may win.

Both apps are excellent on-device dictation tools. The differences come down to: (1) free vs $249, (2) 0 GB RAM vs ~1.5 GB, (3) tuned cleanup vs generic LLM polish, (4) per-app tone matching vs single cleanup style. Dollop wins on all four — but only on Apple Silicon Macs running macOS 26.

If that hardware requirement doesn't fit, Superwhisper remains the most mature alternative. For Mac users who do meet the requirement, the math is clear.