HitsurokuAI voice input workflow for MacAI voice input for work
Hitsuroku
Speak, and get text you can actually use for work.
Your colleague's name misspelled. Your project name garbled. Technical terms that come out as nonsense. Hitsuroku fixes all of this — just describe who you are, who you work with, and what you do, and speech becomes work-ready text from the first output.
Apple Intelligence can cover a basic cleanup flow, and local LLMs or CLI integrations can take formatting further when you need work-ready output with fewer edits.
Download for MacWhy it works
Fewer corrections from the start
General speech-to-text stops at transcription. Hitsuroku adds your business context so output arrives closer to what you'd actually write.
Key proof point
Context
Describe who you are and what you do
Name your colleagues and projects
Preset switching for different jobs
Write a few sentences about yourself in natural language: your name, your role, who you work with, what projects you handle, what technologies you use. That's it. The AI uses this description to get names, terms, and tone right — without you listing every word individually.
Names come out right because the AI knows who they are
"My manager is Siobhan O'Brien" means the AI spells her name correctly — not "Shavon" or "Shivon." It knows who she is, so it uses the right spelling only when you're talking about her.
Your role pulls in the right terms automatically
Write "I'm a site reliability engineer" and the AI already handles Kubernetes, Terraform, Prometheus correctly — without registering each one. Your job description does the heavy lifting.
Transcription Engine Choice
Choose Apple Speech Analyzer, Apple Speech Recognition, or WhisperKit for the speech layer. WhisperKit runs fully on-device when you want a local-only starting point.
Data Minimization
Your text is not sent to Hitsuroku servers. Server communication stays within startup verification and usage count data for account management, while choosing an API formatting provider sends text to that provider.
Formatting Provider Flexibility
Pick local options like Ollama, LM Studio, and Apple Intelligence for private formatting, bring your own CLI tools, or connect through Gemini API depending on how far you want cleanup to go.
Auto submit
Finish formatting, then confirm in the original app
LLM cleanup can take a moment. If you switch away while Hitsuroku is formatting, return to the original app and choose how the text should be confirmed. When formatting ends, Hitsuroku can send that confirm key for you.
Submit keys
This helps avoid the common case where you already dictated the message but later forget to press submit.
Wait without staying on the same screen
You can move to another app while formatting runs instead of staring at the input field.
Return to the destination and pick the key
Go back to the original app and choose Enter, Shift+Enter, or Command+Enter depending on how that field submits.
Send right after formatting ends
When the LLM result is ready, Hitsuroku confirms with the selected key so the drafted text goes out as-is.
Workflow
Build one work-ready pipeline from speech to send
Combine transcription, cleanup, translation, and submit actions so the result fits the field you are writing into, not just a transcript box.
Speech to polished, ready-to-send messages
Dictate with your colleague's name, your project name, your domain jargon — and get clean text with every proper noun spelled correctly. No more fixing names after dictation.
Technical notes with the right terms
Write "I'm a backend engineer using Event Sourcing" in your Context, and the AI handles CQRS, aggregates, and projections correctly — without listing each term. Your role description does the work.
Rough thoughts into ready-to-send text
Dictate rough ideas, and Context fills in the correct name spellings and formatting. The AI knows Nguyen is your teammate and O'Brien is your manager — each name spelled right every time.
Skip formatting when you don't need it
While the AI is still formatting, you can see the raw transcription. If it's already good enough — confirm it instantly without waiting. You choose whether each input needs full formatting or just the transcript.
Profiles
Bring work-specific input modes forward
Each profile combines a Context, transcription engine, and formatting preset into one shortcut. Switch profiles when your vocabulary changes — not just your tone.
Development profile
Context has your engineering team and tech stack. Siobhan O'Brien, Priyanka Krishnamurthy, Sekiban, Azure Container Apps — all spelled right every time.
Translation profile
Same development Context, but with translation turned on. Speak in one language, get polished output in another — with every name and term intact.
Photography profile
Switch to a completely different world. Now the AI knows Xiaowei Zhang and Anaïs Beaumont-Leclerc — not your engineering team. Your vocabulary changes, your profile changes.
Hotkey
Choose your hotkey mode
Two modes to match your permission preferences. Both trigger voice input from any app.
Full Mode
Requires Input Monitoring access. Supports modifier-only shortcuts and native push-to-talk hold behavior.
Limited Mode
No additional system access needed. Supports key-combination shortcuts only, keeping a smaller permission footprint.
Plans
Payment Plan Details Coming Soon
Paid plans with higher daily quotas are being prepared. Hitsuroku Plus includes a free 30-day trial — cancel anytime, and you keep access until the trial period ends.
Plus plan trial
Start with a 30-day free trial. If you cancel during the trial or any billing period, your plan stays active until the end of that period — no immediate cut-off.
Important
Starting April 16, 2026, the free account quota will return to 5 uses per day. If you need more, paid plans will be available.
Get started
Sign in on the web, then set up Context in the Mac app.
Web is for account, usage, and Mac connection management. Context presets, terminology, and writing rules live only in the Mac app and are not sent to the server.
Create your account
Set up your web account and sign in once.
Add Context in the Mac app
Open the Mac app and save your names, terminology, and writing rules there. Those settings stay on that Mac.
Connect your Mac
Link the app to your account so usage and connected-Mac status stay in sync.
Ready to begin?
Create an account or sign in to review usage and connect your Mac. Add Context later inside the Mac app.
Use the account page for sign-in, usage, and connected-Mac management. Add Context only in the Mac app.