日本語 Try free

LIVE CUE SYSTEM

Speak, and words take shape.

Your live-stream voice, set as typography in real time. Even when you’re talking on your own, the screen nods along — a cue card that makes talking a little more fun.

  • No sign-up
  • No install
  • Desktop & phone
  • Data stays on your device
ON AIR 00:00:00;00 Compose · Night
TYPE
LAYOUT
FIG.1 — The real display engine running a demo script (no audio)
¥0Running cost of the free version
1secFrom your voice to words on screen
7faces5 standard + 2 premium typefaces
3minFrom opening the page to talking

01FEATURES

Your words, typeset.

Instead of streaming a raw transcript, KANPE picks out the keywords and sets them on screen with real typographic care. Designed around Japanese script — and it works in English, too.

01

Kanji stand tall, kana step back

Kanji are set bold and large; kana endings shrink to about 80% and get lighter. Katakana go a touch larger because they read small. Every script gets its own treatment.

02

Say it again, and it grows

Each time you repeat a word it gets bigger and bolder, so the topic of the moment is obvious at a glance.

03

A small reward for new words

The first time a standout word comes up in a stream, it arrives in the accent color with an underline — at most one every five seconds, so it stays special.

04

No “um”s on screen

Fillers like “um”, “uh” and “you know” are kept off the screen and out of the transcript. Stumble all you like — the screen stays clean.

05

Pauses leave an afterglow

Talk a lot and the screen gets busy; go quiet and words linger, slowly fading. The screen breathes with the rhythm of your talk.

06

Vertical type and captions

Mincho themes set short kanji words vertically for a Japanese-poster feel. Your latest sentence can run as a caption line that wraps at natural phrase breaks.

02PLUS

Put on a show.

Two effects that come with Plus and up. Free users can try them for a day.

VJ MODE

Words that glow in dots.

VJ mode draws what you say as dot art. Your latest word lights up big in the center while earlier ones shrink and drift around it. New words send ripples across the screen, and the equalizers move with your voice.

  • LED
  • Neon
  • Pixel

SPEAKERS

See who said what.

For collabs and interviews, give each speaker a name label and a color. Switch with the buttons or keys 1–4 — or, on desktop, let KANPE switch automatically by voice pitch. Names carry over into the transcript.

  • Up to 4 people
  • Names & colors
  • Names in TXT & SRT

03HOW IT WORKS

Up and running in under three minutes.

No account, no install. Open it in your browser and start talking.

  1. Open it in your browser

    Chrome or Edge on desktop; Safari on iPhone or Chrome on Android. That’s it.

  2. Allow the mic and talk

    Press “Start talking” and keywords are picked up and set on screen as you speak.

  3. A cue card, or part of the show

    Put it on your own monitor as a cue card, or overlay it in OBS with a transparent background so viewers see it too.

  • Solo chat streamsTopics you can see keep the talk going
  • Game streamsA ticker band that stays out of the way
  • Webinars & talksA cue card so nothing gets forgotten
  • Study sessionsSave the transcript as SRT afterwards

04PRICING

Free, for as long as you like.

Just three plans, designed around streaming two hours, three times a week. Start free; pick Plus for looks and effects, or Pro for accuracy and automation. Prices are in Japanese yen and include tax.

Free

For trying it out. No time limits.

¥0free forever

  • Unlimited browser speech recognition (Chrome, Edge and Safari on desktop and phone)
  • 5 standard typefaces, 3 layouts
  • Cue view and OBS overlay
  • Keywords and blocked words
  • Transcript export as TXT / SRT
  • A one-day (24-hour) trial of the paid looks and effects — premium typefaces, background images, VJ mode, speaker colors and more

A small “KANPE” mark appears in the corner of the display

Start free

Plus BETA

Make the look and effects your own — affordable because they cost us nothing to run

¥480/ month

  • VJ mode: words drawn in dot art (LED, Neon, Pixel)
  • Speaker labels and colors: up to 4 people, switch with buttons or keys, or automatically by voice on desktop
  • All premium typefaces (“Brush”, “Dot” and new ones every month)
  • Any accent color, your own background image, or your camera as the background
  • Remove the “KANPE” mark from the display
  • Unlimited browser speech recognition

Renews automatically. Cancel any time and keep access until the end of the period.

RECOMMENDED

Pro SOON

Everything, for people who stream every week

¥2,980/ month

  • High-accuracy cloud recognition, 20 hours a month
    Enough for three 2-hour streams a week
  • About ¥149 per hour; extra hours ¥200 each
  • Automatic speaker separation and coloring (cloud diarization)
  • Your keywords feed into recognition, so names and jargon come out right
  • AI picks out key points and ideas as you talk
  • A summary after each stream (key points and topic flow)
  • Everything in Plus

Or ¥29,800 / year (two months free)

Payment happens on Stripe’s page — your card details never pass through KANPE. Please read the Terms, Legal notice and Privacy policy before subscribing.

High-accuracy hours count only the time you’re actually talking (a 2-hour stream is roughly 1.2–1.4 hours). Free and Plus use your browser’s built-in speech recognition, so audio is processed by Google in Chrome, Microsoft in Edge and Apple in Safari — never by KANPE. Get in touch about team plans.

05PRIVACY

What you say stays on your device.

KANPE has no server that receives your transcripts or settings. Data moves as little as possible, so you can stream with peace of mind.

All traffic is encrypted (https), and the pages carry no ads, analytics tags or third-party scripts. Pro’s cloud recognition (coming soon) will send audio only to be transcribed and is designed so nothing is stored on KANPE’s servers.

06FAQ

Questions

Is it really free?

Yes. The free version runs entirely on your browser’s built-in speech recognition and keyword extraction in the page, so there’s nothing to pay and no time limit.

Which browsers work? Does it work on my phone?

Voice input works in Chrome and Edge on desktop, Chrome on Android and Safari on iPhone. Allow microphone access when asked. On iPhone, Siri (or Dictation) needs to be turned on in Settings. In-app browsers such as those in LINE or X aren’t supported — reopen the page in Safari or Chrome. The display views alone work in almost any browser, including OBS Browser Sources.

How do I use it with OBS?

Recognition runs in the console in Chrome and the results are sent to OBS. With OBS’s built-in WebSocket server (OBS 28 and later) you can overlay it with a transparent background. For a quick start, you can also capture the display window with OBS’s Window Capture.

Is what I say stored anywhere?

Nothing is stored on KANPE’s servers. Transcripts, settings and background images live only in your own browser on this device, and you can delete them with the Clear button or by clearing site data. Only the step that turns audio into text is handled by your browser vendor’s service (Google in Chrome, Microsoft in Edge, Apple in Safari). For confidential topics, blocked words keep chosen terms off the screen and out of captions.

How do payment and cancellation work?

You pay on Stripe’s page by card or other methods — card details never pass through KANPE. Choose monthly or yearly; plans renew automatically. Cancel any time from “Your plan” → “Billing, cancel, receipts” in the console, and keep access until the end of the period (no prorated refunds). To use your plan on another device, restore it with the license code shown under “Your plan”.

What’s the difference between Plus and Pro?

Plus (¥480/mo) bundles the looks and effects — VJ mode, speaker colors, premium typefaces, background images — and keeps unlimited browser speech recognition. Pro (¥2,980/mo) adds 20 hours a month of high-accuracy cloud recognition, enough for three 2-hour streams a week, plus automatic speaker separation, AI key points and post-stream summaries.

Can it recognize English speech?

Yes. Choose English as the speech language in the console (it’s the default when you open the English version). Keyword extraction and filler removal work in English as well as Japanese.

STAND BY

Ready? Start talking.

Start free in your browser

No sign-up, no install — on desktop or phone