Question 1

How accurate is the transcription?

Accepted Answer

Clear Japanese speech is transcribed with high accuracy. Speaker separation is supported, so conversational audio comes out organized by who said what and when. Specialized terms can be fixed right in the editor.

Question 2

What audio file formats are supported?

Accepted Answer

MP3, WAV, M4A, WEBM, FLAC, and MP4. Files up to 2 hours and 500MB can be uploaded — phone voice memos and online meeting recordings work as-is.

Question 3

How much can I use for free?

Accepted Answer

No credit card required: 30 transcription minutes and 5,000 TTS characters every month. TXT/SRT/VTT export works on the free plan too. Before signing up, you can try up to 90 seconds right on the top page.

Question 4

Is my uploaded audio safe?

Accepted Answer

Audio and transcripts are stored only under your account and never shown to other users. They are not used as AI training data. You can delete your data at any time.

Question 5

Can I turn transcripts back into audio?

Accepted Answer

Yes. Paste or send the edited text to Speech (/speech) to generate natural narration (MP3) — useful for narration, learning materials, and podcast drafts.

Question 6

Can I rename "Speaker 1" to real names?

Accepted Answer

Yes. Renaming a speaker in the transcript editor applies to the entire transcript — ready for minutes and interview drafts.

Question 7

What export formats are available?

Accepted Answer

Speaker-labeled TXT (for minutes/drafts), SRT and VTT (for video subtitles), and clipboard copy. All formats are available on the free plan.

Question 8

How long does processing take?

Accepted Answer

Roughly a few minutes per hour of audio (varies with load). The list status updates automatically, so you can do other work while it runs.

Question 9

Why are web ad CPMs continuing to decline?

Accepted Answer

The main drivers are ad inventory oversupply, and declining targeting accuracy due to third-party cookie restrictions. Additionally, iOS ITP (Intelligent Tracking Prevention) restrictions and the growing adoption of ad blockers like AdBlock are contributing to the decline. Digital Applied 2026 Display Benchmarks reports GDN's average CPM sits at $3.12—cheap for advertisers, but a tough reality for publishers who can only sell at those prices.

Question 10

What are zero-click searches, and how do they affect publishers?

Accepted Answer

A zero-click search is when a user's question is answered directly on the Google results page, so they never click through to a website. Roughly 60% of Google searches now end without a click (The Digital Bloom 2025 Report). Media outlets heavy on how-to articles and fact-checking content are hit hardest. The Reuters Institute predicts search traffic will drop nearly 40% over the next three years.

Question 11

How are AI Overviews impacting media traffic?

Accepted Answer

According to Seer Interactive's research (June 2024 – September 2025), queries that trigger AI Overviews see a 61% drop in organic CTR and a 68% drop in paid ad CTR. At the PIANO Japan user conference, publishers reported 20–30% pageview declines. Informational and news-oriented content is feeling the impact most acutely.

Question 12

How should media companies rethink their revenue models?

Accepted Answer

Moving away from pure ad-dependency is now urgent. Third-party cookie restrictions have already cut ad revenue by an average of 15%. As PIANO Japan's president noted at a Kodansha C Station panel, "hybrid monetization"—subscriptions, events, e-commerce—is becoming the norm. Text alone struggles to capture user attention; diversification into video and audio is increasingly unavoidable.

Question 13

What is the "shift away from text," and how does audio fit in?

Accepted Answer

Younger audiences in particular are moving from text to video and audio consumption. Traffic from TikTok, Instagram, and YouTube is becoming the primary discovery channel. Audio reaches users who "don't have time to read"—during commutes, chores, or any situation where looking at a screen isn't practical. The growth of the podcast market (roughly 460 million listeners as of 2025) is reinforcing this trend.

Question 14

How does audio affect session duration on web media?

Accepted Answer

The average session duration for a text article is measured in seconds, but users who hit play on audio stay for minutes. PUBVOICE's data shows a significant drop in 0–10 second bounce rates. Longer sessions translate directly to higher RPM (revenue per mille) and more repeat visitors. More ad impressions also mean improved ad revenue.

Question 15

Why is everyone talking about media companies building their own apps?

Accepted Answer

The underlying concern is over-reliance on aggregation platforms like Yahoo! News and SmartNews. When 30–50% of your traffic comes from a third-party platform you don't control, that's a fragile position. A dedicated app unlocks push notifications, first-party data collection, and subscription revenue. Japan's new smartphone legislation (enacted December 2025) liberalized external payment options, giving publishers more flexibility in their app strategies.

Question 16

What is GEO (Generative Engine Optimization), and how is it different from SEO?

Accepted Answer

GEO is about being selected by AI as a "trusted source" worth citing. Traditional SEO aims for higher rankings on search engine results pages; GEO aims to get your brand surfaced inside AI-generated answers. PIANO Japan's president has pointed out that "whether AI picks you up" is becoming a new KPI. Media outlets with original primary sources and deep analysis hold an advantage in GEO.

Question 17

Why audio and not video?

Accepted Answer

Video production is expensive and time-consuming. Audio can be AI-generated directly from text, turning existing articles into a new format with near-zero marginal cost. While the average CTV ad is watched for just 22 seconds, audio articles sustain attention for minutes. The growing installed base of smart speakers and wireless earbuds has built a ready audience, and the podcast market (roughly 460 million listeners as of 2025) continues to expand.

Question 18

What is PUBVOICE?

Accepted Answer

PUBVOICE is a SaaS platform that automatically converts web articles into AI-generated audio. Add a "play button" to your articles and readers can listen instead of read. Content is pulled automatically via RSS feed or embedded directly with our SDK. Available since March 2026, you can start free on the Free plan.

Question 19

How does PUBVOICE improve revenue?

Accepted Answer

It works through multiple channels: longer session durations mean more ad impressions, higher repeat-visitor rates, additional ad placements inside the audio player, and subscription-ready infrastructure. When session duration goes up, RPM improves along with it.

Question 20

Is the SDK hard to embed?

Accepted Answer

You just add two things to your HTML: the script tag and a div tag where you want the player to appear. We support WordPress, Next.js, static HTML, and other major platforms.

Question 21

How does automatic RSS feed ingestion work?

Accepted Answer

Register your media outlet's RSS feed URL, and audio is generated automatically whenever a new article is published. The recent articles available in the feed at registration are also converted automatically, and older articles can be converted by submitting their URLs from the dashboard.

Question 22

How is the audio quality? Does it sound natural?

Accepted Answer

We use state-of-the-art AI text-to-speech technology with natural intonation and pacing that comes close to a human narrator. Multiple voice characters are available to choose from.

Question 23

What are the limitations of the Free plan?

Accepted Answer

The Free plan allows a set number of articles to be converted to audio each month. Check the dashboard for current limits. If you need to audio-enable more articles, paid plans start at 1,480 JPY/month.

Question 24

Which languages are supported?

Accepted Answer

Japanese, English, Spanish, and Portuguese are supported. Additional languages are on the roadmap.

Question 25

Will it break my existing site design?

Accepted Answer

The player is fully responsive and designed to blend naturally into your existing layout. Custom CSS adjustments are also supported.

Question 26

Can I track analytics?

Accepted Answer

Audio-specific metrics—play count, completion rate, drop-off points—are available in the dashboard. Google Analytics integration is on the roadmap.

Question 27

Can I use it across multiple media sites?

Accepted Answer

Yes. Multi-site management is supported. Media groups and companies running multiple brands can manage everything from a single account.

Question 28

How long does setup take?

Accepted Answer

Embedding the SDK tag takes just a few minutes. With RSS integration, you can go from feed URL registration to your first generated audio in about 30 minutes.

Question 29

Which platforms are supported?

Accepted Answer

WordPress, Next.js, Nuxt.js, and static HTML sites are all supported. Any environment that runs JavaScript will work.

Question 30

How long does audio generation take?

Accepted Answer

A single article typically takes 1–3 minutes to convert. Even long-form articles finish within 5 minutes. With RSS integration, generation starts automatically in the background as soon as an article is published.

Question 31

Are there API rate limits?

Accepted Answer

The Free plan has a daily cap on the number of audio generations. Check the dashboard for the current limit. Enterprise plans with relaxed limits are on the way.

Question 32

Is it secure? What about SSL?

Accepted Answer

All communication is encrypted via HTTPS. The SDK is loaded through a secure CDN.

Question 33

Can I create a custom voice (voice clone)?

Accepted Answer

Yes. You can train a clone voice using samples of your own narrator or presenter. Providing sample audio is required, but the result is a brand-specific voice that reads your articles.

FAQ & Guides

Transcription & Voice Generation

Publisher Features: Industry (Legacy)

Publisher Features: Product (Legacy)

Publisher Features: Technical (Legacy)

Ready to try it for free?