Input
Input 0 is a local-first macOS dictation app: hold Option+Space, speak, and polished text lands in any app.
If you live on an Apple Silicon Mac and care where your audio goes, Input 0 is the dictation app we'd install first — local Metal-accelerated transcription plus pronunciation-aware term correction ("rui en te" to "React") is a genuinely useful combination for coders and writers. The November 2025 addition of SenseVoice and FireRedASR sharpens its CJK story against a Whisper-only rival. Budget an evening for LLM API setup, and note the license: it's open-sourced under CC BY-NC 4.0, so commercial use isn't covered. Want zero setup instead? Look at a hosted service.
Verified 3d ago · liveness 63/100 · cite: rightaichoice.com/tools/input
- Developers dictating code, commit messages and technical docs
- Writers and creators who want dictation faster than typing
- Privacy-sensitive users who refuse to send audio to cloud servers
- Multilingual speakers needing CJK-focused 99+ language coverage
- Windows or Linux users who need cross-platform dictation
- Anyone wanting a zero-config app with no LLM API key to set up
- Businesses planning commercial use, given the CC BY-NC 4.0 license
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Input 0 if you're on Windows or Linux, use an Intel Mac, need commercial licensing, or want dictation that works without wiring up and paying for your own LLM API key.
The app itself is free, but text polishing only works with your own LLM API key — you pay that provider separately for every dictation you polish.
Input 0's app is free and open source, but the running cost sits with the LLM API you supply for text polishing — that's a per-token bill from whichever provider you configure. For a solo developer or writer on an Apple Silicon Mac, that's typically cheaper than a per-seat cloud dictation subscription, but you're trading subscription predictability for metered API spend.
In short
Input — Input 0 is a local-first macOS dictation app: hold Option+Space, speak, and polished text lands in any app. Best for Developers dictating code, commit messages and technical docs, Writers and creators who want dictation faster than typing, Privacy-sensitive users who refuse to send audio to cloud servers. Free to use.
What's new in Input
Checked 3 days agoAcross the latest 3 updates: 2 feature updates and 1 launch.
Added SenseVoice and FireRedASR engines
Input 0 now includes SenseVoice for improved CJK performance and FireRedASR for Chinese with a reported ~2% character error rate, expanding language support.
Smart model recommendation per language
The app now automatically selects the best AI engine for your language, simplifying setup and improving accuracy without manual tweaking.
Open-sourced under CC BY-NC 4.0
Input 0's code is now available on GitHub, allowing community contributions and transparency for privacy-conscious users.
What people actually say about Input — is it worth it?
We scanned public community sources for Input on Sep 9, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Only 4 of the posts we fetched could be positively tied to Input. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.
Viability Score
How well maintained and how widely used is Input? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Global Option+Space dictation hotkey, works in any macOS app
- On-device speech-to-text accelerated by Apple Silicon Metal GPU
- LLM-powered polishing that fixes grammar and removes filler words
- Auto-pastes transcribed text into the app currently in focus
- Corrects 40+ technical terms by pronunciation ('rui en te' → 'React')
- 10-message context window tracks phrasing across a conversation
- Custom vocabulary for names, brands, and domain jargon
- Auto-learns from your corrections via LLM validation
- 6 AI engines and 12 models covering 99+ languages
- Whisper, SenseVoice, Paraformer, Moonshine, FireRedASR, Zipformer CTC engines
- Smart model recommendation selects the best engine per language
- Download only the models you need
- Voice audio never leaves the Mac; no accounts, clouds, or tracking
- Subtle glass overlay UI designed to preserve focus
- Open source under CC BY-NC 4.0
About Input
Input 0 turns an Apple Silicon Mac into a dictation machine. Hold Option+Space anywhere in macOS, speak, and release — the text is transcribed and injected into whatever app has focus, whether that's Slack, VS Code, or a browser field. There is no window to switch to; the interface is a subtle glass overlay that stays out of the way. The transcription itself runs on-device. Input 0 is written in Rust and accelerates speech-to-text with Apple Silicon's Metal GPU, so your audio never leaves the Mac. The polish step is different: it connects to your own LLM API to strip filler words, fix grammar, and correct 40+ technical terms by pronunciation, so "rui en te" comes out as "React." A 10-message context window lets the model track your phrasing across a conversation. Language coverage is the standout spec. Six AI engines and 12 models span 99+ languages: Whisper for broad multilingual work, SenseVoice tuned for CJK, Paraformer for Chinese and Cantonese, Moonshine for faster English processing, and FireRedASR for Chinese at a reported ~2% character error rate. A September 2025 update added automatic per-language engine recommendation, and in November 2025 SenseVoice and FireRedASR joined the engine lineup. The code is open source under CC BY-NC 4.0, published to GitHub in July 2025. Compared with cloud dictation services, the pitch is control: no accounts, no clouds, no tracking, and only the transcribed text ever reaches an API key you own. That's also the catch — you set up and pay for that key, which puts Input 0 in developer and power-user territory rather than zero-config consumer hands.
Behind the Verdict
Input 0 solves a specific and often-ignored problem: the gap between how fast you talk and how fast you type, without handing your microphone to a cloud provider. The architecture is the product. Speech-to-text runs locally in a Rust binary using Apple Silicon's Metal GPU, so the audio stream never leaves the machine. Everything after that — filling out filler words, fixing grammar, correcting technical vocabulary by pronunciation — is delegated to an LLM API key you supply. That split is honest and sensible: the part that must be private is private, and the part that benefits from a frontier model is swappable. The engine lineup is unusually broad for a single app. Six engines and 12 models cover 99+ languages, with Whisper handling general multilingual work, SenseVoice tuned for CJK, Paraformer for Chinese and Cantonese, Moonshine pitched at 5× faster English, FireRedASR claiming ~2% character error rate on Chinese, and Zipformer CTC as a lightweight offline option. The September 2025 smart-recommendation feature picks an engine per language automatically, which removes the most obvious way a newcomer could pick wrong. You download only the models you need, so a monolingual English user isn't carrying Chinese acoustic models around. The day-to-day details are thought through: a global Option+Space hotkey, auto-paste into the focused app, custom vocabulary for names and jargon, auto-learning from your corrections via LLM validation, and a 10-message context window so the model follows your phrasing across a conversation. The glass overlay is deliberately minimal. The trade-offs are real. This is macOS-on-Apple-Silicon only. Polishing quality is only as good as the LLM key you wire up — if you skip that step or pick a weak model, you get raw transcription. The CC BY-NC 4.0 license means commercial use isn't covered, which rules it out for teams that would otherwise love it. And there's no shared or collaborative transcription; it's a personal tool. Compared with a hosted service that works the moment you sign in, Input 0 asks for more setup in exchange for control over where your voice data goes.
Researching Input? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Input actually fits — and what changes day-one when you adopt it.
Holding Option+Space in VS Code to dictate a commit message, then speaking a code comment. Input 0 transcribes locally and its LLM polish corrects 'rui en te' to 'React' before pasting straight into the editor.
Outcome: Commit messages and comments land in the right file without switching windows or cleaning up filler words by hand.
Drafting an article in a browser-based editor by voice. Audio is transcribed on-device via Metal GPU; only the finished text goes out to the writer's own LLM API key for polishing.
Outcome: Drafting speed approaches speaking speed while raw audio never leaves the laptop.
Switching between English and Chinese in Slack. Smart recommendation picks an engine per language — SenseVoice or FireRedASR for Chinese — and only the needed models are downloaded.
Outcome: One hotkey handles both languages without manually reconfiguring the engine each time.
Use Cases
- Dictate emails and Slack messages without touching the keyboard
- Transcribe voice notes directly into note-taking apps
- Speak code snippets, commit messages, and technical docs into VS Code
- Fill out forms and search queries by voice
- Generate polished meeting notes from spoken summaries
- Dictate in Chinese, Japanese or Korean using SenseVoice or Paraformer
Models Under the Hood
as of 2026-10-17
Limitations
- Requires macOS on Apple Silicon, since transcription runs locally via Metal GPU.
- Text polishing depends on connecting your own LLM API, so output quality is tied to the external model you configure — skip that step and you get raw transcription.
- Only available for macOS; there's no Windows, Linux, mobile or web build.
- Open source under CC BY-NC 4.0, so commercial use isn't covered by the license.
as of 2026-10-08
Verification history
We have re-verified Input 10 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
Showing the 6 most recent of 10 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Input tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0
Ideal for
Apple Silicon Mac users who want local dictation and are willing to bring their own LLM API key for text polishing.
What this tier adds
Starting tier — the app is free, with all 6 engines and 12 models, the Option+Space hotkey, and auto-paste included.
Where the pricing makes sense
The company stage and team size where Input's pricing actually pencils out — and where peers do it cheaper.
Input 0's app is free and open source, but the running cost sits with the LLM API you supply for text polishing — that's a per-token bill from whichever provider you configure. For a solo developer or writer on an Apple Silicon Mac, that's typically cheaper than a per-seat cloud dictation subscription, but you're trading subscription predictability for metered API spend.
Setup time & first value
How long it actually takes to get something useful out of Input — broken out by persona, not the marketing-page minute.
For a developer: about 15-20 minutes — install the app, download the model for your language, and wire up an LLM API key for polishing. For a non-technical writer: closer to an evening, because the LLM key setup and picking an engine are the friction points. Once configured, daily use is instant: hold Option+Space and talk.
Switching to or from Input
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From macOS built-in dictation: install Input 0, keep the same Option+Space muscle memory, and gain on-device Metal transcription plus LLM polishing.
- →From a cloud dictation service: install Input 0, download your language model, and point it at your own LLM API key to move audio processing back on-device.
- →From typing-only workflows: add the Option+Space hotkey to your existing apps — Slack, VS Code, browser — no per-app setup required.
- ↗To a hosted dictation service: expect to lose the local-only audio guarantee in exchange for zero API-key setup.
- ↗To macOS built-in dictation: drop the LLM polish and multilingual engine choice, and get back a fully zero-config experience.
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Input”, and we withheld 6: 6 could not be judged, because “Input” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Input.
Official links
Tools that pair well with Input
Common stack mates teams adopt alongside Input, with the specific reason each pairing earns its keep.
Wispr Flow
Wispr Flow turns spoken rambling into polished text in any app, with AI voice dictation on Mac, Windows, iOS, and Android.
Lemon
System-wide macOS and Windows voice tool that turns rambling speech into structured prompts and clean dictation in any text field.
Vocalinux
Free AGPL-3.0 Linux voice dictation: hold Right Alt and local whisper.cpp transcribes into any focused app
Featured Head-to-Head Comparisons
Input vs Gem
Gem and Input serve entirely different needs: Gem is an enterprise-grade hiring platform for teams, while Input is a free, privacy-focused dictation tool for individual macOS users. Your choice depends on whether you need to automate recruitment workflows (Gem) or speed up typing with voice (Input). They are not direct competitors.
Input vs Guesty
If you manage a vacation rental portfolio and need AI-driven automation across bookings, messaging, and finance, Guesty is the clear choice. If you're a macOS power user seeking to dictate text hands-free with full privacy, Input0 is perfect. They serve completely different needs—choose based on your domain.
Input vs Poke Interaction Co
Choose Input if you need fast, private voice dictation on macOS and already have an LLM API key; choose Poke if you want a chat-based AI assistant that manages email, calendar, and automations across messaging apps. They solve different problems: one is a voice input tool, the other is a personal assistant. For most tasks beyond dictation, Poke is more versatile.
Alternatives to Input
View all
Wispr Flow
Wispr Flow turns spoken rambling into polished text in any app, with AI voice dictation on Mac, Windows, iOS, and Android.
Frequently Asked Questions
Best-of guides
Topics
Used Input? Help shape our editorial sentiment research.