Dictation,
on Apple Intelligence.
When Apple shipped Apple Intelligence with macOS 26, it quietly opened a door no one else has walked through yet: third-party apps can now use Apple Foundation Models — the on-device LLM Apple maintains as part of the OS — for any text task, including dictation cleanup. Dollop is currently the only Mac dictation app that does this, which is why it adds 0 GB of RAM, costs nothing, and runs entirely on-device. Here's the full picture.
Apple shipped a 3-billion-parameter on-device LLM as part of macOS 26 and iOS 19. The Foundation Models framework lets any app use it — for free, with no cloud routing, no per-request cost, and crucially, no separate model download. Here's what that unlocks for dictation specifically.
Whisper-based dictation apps load a 700 MB to 3 GB model into memory while running. Apps using Apple Foundation Models don't load anything — Apple's model lives in the OS, kept warm by the system whether or not your dictation app is running. Net new RAM cost from a dictation app: 0 GB.
Apple's foundation model never touches the network — it runs on your Neural Engine. No request goes out, no response comes back. That's a stronger architectural privacy posture than even Whisper-based "on-device" tools, which still ship a model you have to trust the source of.
Generic LLMs over-edit dictation — they smooth tone in ways the speaker didn't intend or hallucinate phrases that weren't said. Dollop trains a LoRA adapter specifically on dictation cleanup pairs (raw → cleaned), which behaves more conservatively than generic GPT polish. The adapter weighs ~150 MB and loads on top of Apple's 3B parameter model.
Cloud-based dictation tools (Wispr Flow, Cluely) have real per-API-call costs that flow into your subscription. Foundation Models has no per-request cost because it runs on hardware you already own. That's why Dollop can be free in a way that cloud apps fundamentally can't.
The hard requirement: Apple Foundation Models needs Apple Silicon and macOS 26 (Tahoe) or later. Most Macs sold since 2021 qualify. Intel Macs and pre-26 macOS versions don't get access to the foundation model and can't run Dollop — Voibe or VoiceInk cover that gap.
Most comparison pages list ten features. Three of them matter when you're actually switching.
- Where does the audio go?
- On-device means your voice and the transcript stay on the Mac you're typing on. Cloud means audio is uploaded, transcribed remotely, and the text is downloaded back — usually with no way to know which providers see it on the way through. This is load-bearing for anyone touching client work, regulated content, or proprietary information.
- How much RAM does it actually take?
- The forgotten metric. "On-device" doesn't mean "free" — Whisper-based tools load 700 MB to 3 GB into memory while running. If you already have Cursor, Chrome, Slack, and a Docker container open, that matters. Apple Foundation Models is the only path right now to real on-device dictation at 0 GB.
- How clean is the cleaned text?
- Raw transcription is a solved problem — Whisper, Parakeet, and Apple's models are all excellent. The differentiator is what happens after: filler removal, punctuation, formatting commands, tone-matching to the active app. A tuned dictation model beats a generic LLM polish step every time.
| Tool | RAM | Audio leaves Mac? | Pricing | Cleanup |
|---|---|---|---|---|
| Dollop | 0 GB | Never | Free | Tuned LoRA |
| Superwhisper | ~1.5 GB | Optional | $249 once | Generic LLM |
| Voibe | ~700 MB | Never | $9.90/mo | Whisper + rules |
| VoiceInk | ~2 GB | Never | Free / open-source | Optional |
| MacWhisper | ~1–3 GB | Never | €64 once | File-based |
| LumeVoice | ~700 MB | Never | Subscription | Whisper |
| Apple Dictation | ~0 GB | Never (Ventura+) | Free (built-in) | None |
RAM figures are approximate, measured at idle with the smallest model loaded. Dollop uses Apple Foundation Models — the OS keeps the model warm independently of the app, so the app itself contributes 0 GB beyond the OS baseline.
Dollop
Dollop is the voice OS I built for the Mac. It does four things, all on-device, all free: dictation (clean text pasted at your cursor in any app), an AI chat overlay (multi-turn, optionally screen-aware) for asking questions or rewriting, Little Overlay (⌃ A — a tiny floating glass pill that hears your selection, sees your screen, and answers in place for the one-shot question that doesn’t need a chat thread), and Meeting Notes (record and transcribe lectures, calls, and meetings with summaries and action items, fully on-device). It runs on Apple Foundation Models, the on-device LLM Apple ships with macOS 26, and ships with a tuned cleanup adapter trained specifically for the way people actually speak. Because Apple already keeps the foundation model warm in the OS, Dollop contributes 0 GB to your RAM footprint. Your audio never leaves the Mac. Completely free.
- Four pillars in one app: dictation, AI chat, Little Overlay, Meeting Notes
- Only Mac dictation app on Apple Foundation Models
- 0 GB RAM cost, the OS hosts the model
- Tuned cleanup adapter, not a generic polish step
- Meeting Notes: record + transcribe + summarize on-device
- Per-app tone (Casual / Formal / Excited)
- Completely free, no subscription, no premium tier
- Apple Silicon + macOS 26 only, no Intel, no Windows
- Apple Intelligence must be enabled
- New product, smaller community than Wispr or Superwhisper
Apple Dictation
Built into macOS Ventura and later. Uses Apple's speech recognition, fully on-device for supported languages on Apple Silicon. Doesn't yet use Foundation Models for cleanup — Apple Dictation is transcription-only, no filler removal, no formatting commands. Dollop adds the Foundation Models cleanup layer on top of the same Apple-native foundation.
Superwhisper
Doesn't use Apple Foundation Models — runs its own Whisper model locally (~1.5 GB RAM). $249 one-time. The mature pick before Dollop existed. Cross-platform (Mac/Windows/iOS) by virtue of not being Apple-exclusive.
Voibe
Whisper-based, $9.90/month. Strong privacy positioning but doesn't tap Apple Foundation Models — still loads ~700 MB of Whisper into memory.
VoiceInk
Open-source Whisper-based dictation. Free. Doesn't use Foundation Models but is auditable and works on older Macs.
Wispr Flow
Cloud-based — sends your audio to remote servers and processes it through OpenAI / Anthropic / Cerebras LLMs. Doesn't use Apple Foundation Models. $15/month.
- Audio leaves your Mac
- $15/mo subscription
- Privacy crisis in early 2026
Hard requirement: Dollop runs on Apple Silicon Macs with macOS 26 or later. Most Macs sold since 2021 qualify. If you're on Intel, that's the constraint to plan around.
What is Apple Foundation Models?
How is Apple Foundation Models different from Apple Intelligence?
Why doesn't Apple Dictation use Apple Foundation Models for cleanup?
Does Dollop work without Apple Intelligence enabled?
What does Dollop actually do?
How much does Dollop cost?
Does Dollop work without an internet connection?
Can Dollop record and transcribe meetings or lectures?
Will Dollop work on Intel Macs?
Can I dictate into Cursor, VS Code, Slack, and other apps?
Will more dictation apps use Apple Foundation Models?
How much faster is dictation on Apple Foundation Models vs Whisper?
Apple Intelligence opened up a category that didn't exist before: third-party apps that ride on Apple's foundation model instead of shipping their own. Dollop is currently the only Mac dictation app doing it — which is why it can be free, on-device, and 0 GB of RAM, all at the same time.
If you're on an Apple Silicon Mac with macOS 26 and Apple Intelligence is on, Dollop is the dictation app that fits the platform best. If not, the older alternatives still work — they just don't get the architectural benefits Apple Foundation Models provides.