AICHE vs SuperWhisper

Short answer: SuperWhisper is the deepest Apple-Silicon-first dictation app in the Whisper-based category, with a real offline path and a dense per-app mode system. AICHE runs on 9 platforms (macOS, Windows, Linux, iPhone, iPad, Apple Watch, Android, Chrome, Obsidian), with REST API access, starts at $3.99/mo billed annually, ships a code-tuned recognition profile, and has a real self-serve team plan. If you live entirely on a Mac and want to transcribe on a plane, SuperWhisper. If you bounce between Linux, Windows, Android, and your phone, or you want to call voice transcription from your own code, AICHE.

What SuperWhisper is selling

SuperWhisper (superwhisper.com) leads with "Turn your voice into polished text" and "Write 3x faster, without lifting a finger." The recurring trio in their marketing is speed, polish, and on-device privacy: "Everything stays on your device," "Works offline, so you can transcribe anytime," and "Supports 100+ languages & dialects." The home page positions the product as system-wide dictation that "works anywhere you can type or paste text, no need to switch apps."

The feature pitch is built around modes. SuperWhisper ships predefined modes that "optimize tone, structure, and formatting" and lets users write their own custom modes with per-app triggers and prompt control. There's a custom vocabulary system for phrases, names, links, and acronyms. Meeting recording with automatic notes is included on the free tier. Pro adds the larger local Whisper model sizes, unlimited custom modes, file and video transcription, translation, and bring-your-own-key support for third-party LLM rewriting (OpenAI, Anthropic, Groq, and similar).

Independent reviewers (Spokenly, Voibe, Product Hunt, May 2026) describe SuperWhisper as the "power-user pick" in the Whisper-based category - flexible, fast on Apple Silicon, accurate, and habit-forming if you commit to the setup. The common counterweights surfaced across those same reviews: dense settings, steep learning curve, a Windows build that Spokenly describes as "less polished than the Mac build" with reported crashes and target-app freezes, default-mode switching that reviewers report doesn't always revert when you change apps, and a feature-request board that Voibe reports has 93.5% of 476 user-submitted requests unaddressed.

The feature comparison

	AICHE	SuperWhisper
Platforms	macOS, Windows, Linux, iPhone, iPad, Apple Watch, Android, Chrome extension, Obsidian plugin, REST API	macOS (v2.14.0 per their changelog), Windows (x64 + ARM64), iOS 18+. No Linux, no Android, no Chrome, no Obsidian (Voibe platform write-up, May 2026).
Free tier	7-day free trial of Pro features, no credit card	Free tier with small local Whisper models, voice-to-text in any app, meeting recording, capped custom modes
Paid floor (annual)	$3.99/mo (Personal annual, $47.99/yr)	$7.08/mo equivalent ($84.99/yr Pro annual)
Paid monthly	$4.99/mo Personal, $9.99/mo Pro	$8.49/mo Pro
Lifetime	Subscription only, no one-time purchase	$249.99 one-time, marketed as "best value"
Team plan	Pro tier includes admin panel, seats, roles, one unified bill	"Custom" Enterprise tier (no self-serve team plan; their own vs-Wispr page concedes team/admin features are not their strength)
Public REST API	Yes, Pro tier. Call AICHE from your own scripts and tools.	No callable API. BYO-key for third-party LLMs (the inverse direction).
Recording model	Toggle hotkey: ⌃+⌥+R (Mac), Ctrl+Alt+R (Win/Linux). Press to start, press to stop.	System-wide dictation via clipboard/paste, configurable shortcuts and per-app triggers.
Custom vocabulary	50 entries, synced across every platform	Yes, taught once and remembered
Code-tuned recognition	Software Development profile (Pro): tuned for code identifiers, CLI flags, library names, kebab-case, snake_case	Not built in. Achievable via custom vocabulary + custom modes (user-configured).
AI cleanup	Built in. Removes filler ("um," "uh," "like"), stutters, repetitions. Auto punctuation and paragraph breaks.	Modes apply tone/structure/formatting rewrites. Cloud rewrite needs BYO key on Pro.
Speed	Sub-100ms cold start. ~3 seconds to transcribe 15 minutes of audio (cloud, Groq).	8-27x realtime on Apple Silicon depending on chip (Spokenly, May 2026). Slower on Intel Mac (where Spokenly reports local performance is poor) and the Windows build, which Voibe characterizes as "available with stability caveats."
Languages (voice input)	99 transcription languages, same engine on every surface	100+ languages and dialects (Whisper engine), with translation to English
Mobile UI localization	28 languages (iPhone, iPad, Apple Watch, Android)	iOS app; UI localization breadth not their headline feature
Apple Watch capture	Yes, record from wrist, syncs to iPhone	No
End-to-end encrypted cloud sync	Yes, opt-in. AES-256-GCM with Argon2id passphrase-derived key. We cannot read your notes.	Local-storage-first model. No equivalent multi-device E2EE sync product.
Offline transcription	No fully offline desktop path. Cloud transcription via Groq, immediate audio purge after processing. Honest about this.	Yes on Apple Silicon. Multiple local Whisper sizes including large-v3 on Pro. Genuinely works on a plane.
Voice into AI coding agents	Voice Code (Pro): pause-aware auto-send into Claude Code, Codex, Cursor, Antigravity	Pastes into any app including IDEs. No dedicated agent-loop workflow.
Privacy posture	Cloud transcription via named provider (Groq), audio purged within seconds, opt-in E2EE sync, no keystroke logging, no clipboard monitoring outside explicit user action	Local-capable on Apple Silicon. Configurable to involve third-party LLMs when user opts in via BYO key. Independent reviewers (Spokenly, Voibe, May 2026) report default on-disk audio retention with no built-in opt-out toggle and plaintext API-key storage for connected providers.
Refund / sample	7-day free trial, no card	30-day no-questions refund, 15 free Pro minutes
Student discount	Not currently advertised	40% off Pro mentioned on pricing page

Where SuperWhisper wins

A few things are honestly theirs, and a comparison article that pretends otherwise isn't worth reading.

The offline path on Apple Silicon is real. If you're on an M-series Mac and need to transcribe while disconnected (flight, train tunnel, on-call rotation in an air-gapped network, paranoid security review of an unreleased product), SuperWhisper actually works. Bundled Whisper model sizes from small through large-v3 run locally. AICHE does not have this. Our desktop transcription streams to Groq, processes server-side, and discards audio within seconds. We are honest about it. If your single hard requirement is "must work offline," SuperWhisper wins this row.

Mode system depth. SuperWhisper's per-app modes with user-written prompts and shell triggers are the deepest configuration surface in this category. If you want one prompt template for Slack, another for git commit messages, and a third for emails to your CTO, and you want each to auto-activate when the right app is in focus, SuperWhisper rewards the time you put into configuring it. AICHE's model is the opposite: one global hotkey, one cleanup pipeline, with a Software Development profile you toggle on. We optimize for "press, speak, get clean text, done." If you want fine-grained mode switching per app, SuperWhisper is built for that.

Bring-your-own-key for cloud LLMs. Pro users with their own OpenAI, Anthropic, or Groq accounts can route the rewrite step through their own billing and their own rate limits. For people who already have a paid API account and want their voice text rewritten by their preferred model, this is a real feature. AICHE doesn't offer BYO key today; cleanup runs through our pipeline.

Lifetime tier exists and is actively sold. $249.99 one-time. If you've calculated that you'll use voice dictation for more than three years and you refuse subscriptions on principle, SuperWhisper has a product for that. AICHE is subscription only, no one-time purchase. Different pricing models with different tradeoffs - subscription tracks ongoing development and infrastructure costs; one-time purchase trades future updates and cloud capacity for a fixed cost.

Apple Silicon performance numbers. Independent reviews (Spokenly, May 2026) report 8-27x realtime transcription on M-series chips. That's local hardware doing the work, not a network round-trip. For very short utterances on a fast Mac, that's faster than any cloud trip can be.

Brand recognition in the Mac power-user community. Product Hunt, MacStories, indie dev Twitter, and the "Whisper-based dictation" discourse all know SuperWhisper. If you're asking your Mac-only friends what they use, SuperWhisper is one of the two or three names you'll hear.

Where AICHE wins

1. Platform coverage (9 platforms vs 3)

SuperWhisper ships macOS, Windows, and iOS. AICHE ships macOS, Windows, Linux, iPhone, iPad, Apple Watch, Android, a Chrome extension, an Obsidian plugin, and a public REST API. That's not a marketing flourish, it's an enumeration:

• macOS - global hotkey ⌃+⌥+R, inserts at cursor
• Windows - global hotkey Ctrl+Alt+R, inserts at cursor
• Linux - global hotkey Ctrl+Alt+R, four package formats: .deb (Debian / Ubuntu / Mint), .rpm (Fedora / RHEL / Rocky), AppImage (any glibc distro), Flatpak (sandboxed, Flathub-ready)
• iPhone (iOS 15+) - tap mic, talk, save, auto-copy to clipboard
• iPad - full tablet layout
• Apple Watch - record from your wrist, syncs to iPhone via WatchConnectivity
• Android - phone, tablet, home-screen one-tap widget
• Chrome extension - voice into any web text field
• Obsidian plugin - voice notes inside Obsidian (Community Plugins listing)
• REST API - call AICHE from your own code

If you run Linux as your daily driver, SuperWhisper has nothing for you. If you carry an Android phone, SuperWhisper has nothing for you. If you live in Obsidian or capture from a Chrome web app, you'd have to add another tool. AICHE is one account, one voice layer, every device you actually use.

The reverse is also worth saying clearly: if you only ever touch a Mac and an iPhone, the platform-breadth argument is weaker for you. SuperWhisper's three-surface coverage is sufficient for that user. The case for AICHE on a pure-Apple stack is the API, the team plan, the dev profile, and the price floor, not the platform count.

2. Real public REST API

AICHE Pro exposes a callable REST API. You send audio, you get text back, you wire it into whatever script, automation, internal tool, or agent loop you're building. This is the single biggest functional gap between the two products if you write code for a living.

SuperWhisper does not expose a callable API. The closest equivalent is bring-your-own-key for third-party LLM providers, which is the inverse direction: the user supplies the key, SuperWhisper consumes it for rewrite steps. You can't call SuperWhisper from your own backend.

Concrete things AICHE's API unlocks: a CLI tool that transcribes voice memos into your project's notes/ directory; a Slack bot that transcribes audio messages in a channel; an internal helpdesk widget that lets non-technical staff dictate tickets; a transcription step inside a larger pipeline you're already running. None of that is buildable on top of SuperWhisper.

3. Real self-serve team plan

AICHE Pro includes an admin panel with seats, roles, and one unified bill. Add a coworker, assign a role, pay one invoice. If your team grows, you add seats.

SuperWhisper has an Enterprise tier with custom pricing and a "contact sales" path. Their own comparison page against Wispr Flow concedes: "for centralized vocabulary management across a large team, [Wispr Flow] offers specialized functionality." If you're three developers and a designer who all want voice dictation under one bill, AICHE has a checkout flow for that; SuperWhisper has an email form.

4. Software Development profile (Pro)

AICHE ships a Software Development recognition profile tuned for code identifiers, CLI flags, kebab-case library names, snake_case function names, and developer jargon. You toggle it on once and the recognizer stops fighting you when you say "use the --no-edit flag" or "in the useEffect hook."

SuperWhisper offers custom vocabulary and custom modes, which a determined developer can configure for the same job. Both are genuinely useful, but they're user-built. The difference is between "we tuned the model for code, turn the switch on" and "build your own prompt and vocab list and test it across the identifiers you use." The first one works the day you install it. The second one works after a Saturday afternoon of setup.

If you live in Cursor, Claude Code, Codex, or Antigravity, AICHE Pro also has Voice Code: pause-aware auto-send into the agent's input box. You speak, you pause, the prompt ships on its own. No Enter, no Send button. SuperWhisper pastes text into any app including IDEs, which works, but the agent-loop pattern (long prompt, pause, auto-send, listen for the next response, dictate the follow-up) isn't a dedicated workflow.

5. Pricing floor

AICHE Personal annual: $3.99/mo equivalent ($47.99/yr). Three devices, all the capture, cleanup, sync, custom vocab, and 99-language transcription you'd actually use. 7-day free trial, no credit card.

AICHE Pro annual: $8.33/mo equivalent ($99.99/yr). Ten devices, plus the Software Development profile, REST API, team plan, priority processing, priority support, and Voice Code for AI agents.

SuperWhisper Pro annual: $7.08/mo equivalent ($84.99/yr). One tier.

Two reads on this:

• If Personal is enough for you: AICHE Personal is roughly 56% the price of SuperWhisper Pro annual, with E2EE sync, custom vocabulary, and 99-language transcription included. AICHE is the cheaper product.
• If you need Pro features (API, team plan, dev profile, Voice Code): AICHE Pro annual is $99.99 vs SuperWhisper Pro annual $84.99. AICHE costs $15 more per year and unlocks things SuperWhisper doesn't offer at any tier.

The $249.99 lifetime tier is a separate calculation. Breakeven against AICHE Pro annual is roughly 30 months; against AICHE Personal annual is roughly 5 years. If you're a heavy user who refuses subscriptions and you've already decided on local-on-Apple-Silicon, the lifetime tier is rational. AICHE is subscription only; the annual tiers are the comparison for most buyers.

See /pricing for the full breakdown including current promotional rates.

6. Speed and quality with honest framing

AICHE: sub-100ms cold start, ~3 seconds to transcribe 15 minutes of audio, cloud transcription via Groq with named provider disclosure and immediate audio purge after processing. Speaking runs ~150 WPM; the bottleneck is your speech, not the pipeline.

SuperWhisper: 8-27x realtime on Apple Silicon (Spokenly, May 2026). Local model on a fast Mac, no network round-trip. Slower on Intel Macs - Spokenly reports "Intel Macs run cloud models but local performance is poor" - and on the Windows build, which Voibe's platform support write-up characterizes as "available with stability caveats" (the macOS build is the flagship experience). On an M-series Mac running Whisper Large v3 locally, a 15-minute recording typically processes in roughly 33-112 seconds (the literal range derived from the Spokenly 8-27x realtime numbers), an order of magnitude slower than AICHE's ~3 seconds on the same input, before any post-processing.

Both can be fast. The question is which kind of fast you want. SuperWhisper is "no network needed, your laptop does it all" fast. AICHE is "Groq returns 15 minutes of audio as text in 3 seconds, regardless of how many cores your laptop has" fast. If you're on a low-power machine, AICHE's cloud path is often faster than a local Whisper pass. If you're on a maxed-out M-series Mac with no internet, SuperWhisper wins.

On quality: raw Whisper output is not finished text. The model regularly produces filler-word artifacts, phantom phrases when audio quality dips (the documented "thanks for watching" / "subscribe" insertions that appear in unrelated recordings), mistranscriptions of proper nouns, and punctuation drift. Local-Whisper apps that ship the model output directly ship those failure modes too.

AICHE runs a multi-stage polish pipeline after Whisper: an empirical hallucination filter built from millions of recordings catches the known phantom phrases, filler and stutter removal cleans "um" / "uh" / false starts, your 50-entry custom vocabulary gets enforced on every pass, the Software Development profile tunes recognition for code-adjacent speech (Pro), and a fast LLM polish through Groq smooths grammar without changing your voice (zero retention, no logging, no training). That whole pipeline runs in the ~3 seconds, included on every tier. SuperWhisper's mode system can do similar rewrites, but the cloud-LLM rewrite step on Pro requires you to bring your own API key, and the empirical hallucination filter / dev-tuned recognition aren't features they ship.

A note on privacy framing

Both products are doing honest things here, in different ways, and the comparison is more useful as a checklist than as an accusation.

SuperWhisper's pitch is local-first: "Everything stays on your device." On Apple Silicon, the default workflow can be fully offline. The privacy policy disavows server-side audio, transcript, and usage data retention. That's a genuine differentiator versus cloud-only competitors, and it's what their power-user community values most.

The architecture is configurable, though. Pro's bring-your-own-key support means the rewrite step can route through third-party LLM providers when the user opts in. Independent reviewers (Spokenly and Voibe, May 2026) report that audio files save to local disk by default with no built-in opt-out toggle, and that connected provider API keys are stored in plaintext JSON rather than the macOS Keychain. None of that is unusual for a power-user Mac app. It's worth knowing if you're comparing on privacy.

AICHE's pitch is cloud-with-deletion, named provider, and opt-in E2EE sync. Specifically:

• Audio is streamed to Groq (named transcription provider), processed in seconds, and discarded. No persistent audio storage. Typical retention: 2-3 seconds during processing, hard cap within 1 second after processing.
• Transcripts live locally on your device by default. Cloud sync is opt-in.
• When you turn sync on, notes are encrypted on your device before they leave it: AES-256-GCM with an Argon2id-derived key from a passphrase you set. We cannot read synced data on our servers.
• Transmission uses modern TLS, with certificate pinning on iOS and Android.
• No global keystroke logging. No clipboard monitoring outside of explicit user action. No active window title transmission. No microphone monitoring outside of explicit recording. No background recording. No shared hardcoded credentials between users.

If you're comparing voice apps on privacy, the questions that matter for any of them are:

1. Is the audio retained on disk after transcription, and is there a toggle?
2. Where are connected-provider API keys stored, and in what format?
3. Does the cleanup or rewrite step involve a third party, and how is that disclosed?
4. Is the always-on listener a walkie-talkie button (you press it to start) or an always-on security camera (it listens until you press stop)?
5. If there's cloud sync, is it end-to-end encrypted, or can the vendor read your notes?

Both products will give you defensible answers. The specifics differ, and the right product depends on which threat model you actually have.

Common questions

Q: I'm on Linux. Which one works for me?

A: AICHE. SuperWhisper has no Linux build, and there's no "coming soon" marker on their download page. AICHE ships four package formats across distro families: .deb (Debian / Ubuntu / Mint), .rpm (Fedora / RHEL / Rocky), AppImage (any glibc distro), and Flatpak (sandboxed, Flathub-ready). Global hotkey is Ctrl+Alt+R, inserts via the standard X/Wayland event path. If you live on Linux, this isn't a close call.

Q: I'm on Mac and I want to dictate offline. Which one?

A: SuperWhisper. Their local Whisper path on Apple Silicon genuinely works on a plane. AICHE's desktop transcription requires internet because the audio streams to Groq. There's a Later queue for crash recovery, but real-time transcription needs a network. If your single hard requirement is offline, SuperWhisper wins.

Q: I dictate prompts into Cursor and Claude Code. Which one handles snake_case and CLI flags better?

A: AICHE Pro, by default. The Software Development profile is recognition tuned for code identifiers, kebab-case flags, snake_case names, and library names. Voice Code adds pause-aware auto-send into Claude Code, Codex, Cursor, and Antigravity. SuperWhisper can be configured for similar results via custom vocabulary plus custom modes, but it's user-built; AICHE's dev profile ships tuned and is a toggle.

Q: I want to call voice transcription from my own code. Is there an API?

A: AICHE has one. Pro tier exposes a public REST API: send audio, get text back, wire it into whatever you're building. SuperWhisper does not expose its own callable API. Their BYO-key feature lets you plug your OpenAI/Anthropic key into the rewrite step, but you can't call SuperWhisper from a backend.

Q: I'm setting up voice for my team of five. Which one has the lower-friction team plan?

A: AICHE. Pro tier includes a self-serve admin panel with seats, roles, and one bill. SuperWhisper's team path is an Enterprise "contact sales" tier with custom pricing, and their own competitor page concedes team/admin features are not their strength.

Q: I carry an Android phone and a Linux laptop. I want one voice tool across both.

A: AICHE. SuperWhisper has no Android app (the top-voted pending request on their public feedback board, with 198 votes per Voibe's May 2026 platform write-up) and no Linux build. AICHE runs natively on both, plus the macOS, Windows, iPhone, iPad, Apple Watch, Chrome, and Obsidian surfaces you'll probably want eventually.

Q: I want a one-time payment instead of a subscription. What's the move?

A: SuperWhisper's $249.99 lifetime tier is built for you. AICHE is subscription only - no one-time purchase option. If "no subscription" is a hard constraint, SuperWhisper has the product for that. If you're flexible, AICHE Personal annual at $3.99/mo is roughly $48 per year and undercuts the breakeven math for most users under a five-year horizon.

Q: I speak Russian, German, or Japanese into the mic. Does either handle it?

A: Both. Both products run on top of Whisper at the engine level, so coverage is essentially the same ~99 languages either way - SuperWhisper markets it as "100+", AICHE quotes the underlying engine's 99. AICHE adds auto-translation to English available everywhere (mobile, desktop, Chrome, Obsidian). Worth a small detail: AICHE's mobile UI is localized to 28 languages (iPhone, iPad, Apple Watch, Android); desktop, Chrome, and Obsidian UI is English only. Voice input works identically across every platform regardless of UI language - a Russian speaker on Linux gets Russian transcription even though the menus are in English.

A few honest tradeoffs on the AICHE side

In the spirit of the "where they win" section above, here are the things AICHE doesn't do, which the SuperWhisper crowd will rightly point out:

• No fully offline desktop path. Cloud transcription needs internet. We're honest about it; if "must work offline" is non-negotiable, SuperWhisper.
• Desktop UI is English only. Mobile UI is localized to 28 languages. Voice input works in 99 languages everywhere.
• Software Development profile is Pro only. Free Personal tier doesn't get it.
• API access is Pro only.
• Apple Watch recording requires iPhone for the WatchConnectivity relay.
• Heavy multi-device users on Personal may hit the 3-device cap. Pro goes to 10.

Result

If you live on a Mac, value an offline path, want the deepest per-app mode system, and are happy to invest a weekend in setup, SuperWhisper is the right tool. They have years of focused Mac development behind them and a real power-user community.

If you bounce between Linux, Windows, Android, and a phone; want to call voice transcription from your own code; need a self-serve team plan; want code-tuned recognition out of the box; or want to dictate into AI coding agents with pause-aware auto-send, AICHE is the right tool. 9 platforms, one account, $3.99/mo billed annually to start.

Try it now: install AICHE on whichever device you're on, press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux), and dictate the next thing you were about to type. Seven-day free trial, no credit card. See pricing.