Why a Publisher Got Kicked Off Apple Books After Launching 29-Language Audiobooks with Auto-Narration — The 5 Quality Standards Ridi and Millie Actually Enforce

In January 2026, an overseas publisher used Apple Books' auto-narration feature to simultaneously launch a bestseller in 29 languages—only to be removed from the platform three weeks later. The reason was simple: "quality standards not met." Mass-produced audiobooks made with free TTS couldn't clear even the platform's baseline quality bar, and a flood of listener refund requests prompted the platform to pull the content directly. While the audiobook market grew to $40.3 billion in 2026, platforms began filtering content not by 'fast and many' but by 'quality worth listening to.'

Key Takeaways

The Audiobook Market Is Growing—So Why Are Platforms Rejecting Content?

The audiobook market is expanding rapidly. As of 2026, the global audiobook market is estimated at $40.3 billion, and the Korean market—centered on Ridi, Millie, and Welaaa—maintains over 20% annual growth. Demand from publishers and platforms to convert web novel IP into audiobooks has surged. Yet as the market grows, platform quality standards are rising in tandem.

Leading global platforms like Apple Books and Audible now clearly specify quality review criteria for audiobook registration. For example, Apple Books requires that AI-narrated audiobooks be explicitly labeled as "AI narration" and reserves the right to reject registration or remove already-published content if quality standards are not met. Repeated submissions, inaccurate language metadata, or bulk submission of similar content can be flagged as "spamming" and result in distribution blocks.

The primary reason platforms tighten quality standards is to protect listener experience. Apple Books officially states that if AI narration does not meet quality criteria, it will not publish the content and will notify the partner. Listener complaints lead to refund requests and platform churn, which in turn drives up subscription cancellation rates. This trend is consistent across all major platforms, domestic and international.

Why Can't Audiobooks Made with Free TTS Sites Clear the Platform Bar?

Publishers face a genuine dilemma. They want to convert web novel IP into audiobooks quickly, but hiring a professional production house is expensive and time-consuming. So many turn to free TTS sites first. Tools like Google TTS, Naver Clova Dubbing, and MS Azure TTS are fast and convenient—paste in text, and you get an audio file in minutes.

The problem is that the output doesn't pass platform quality standards. Free TTS has three major limitations:

Web novel audiobooks often feature heavy dialogue and multiple characters. With free TTS, the male lead, female lead, and supporting cast all speak in the same voice. Listeners can't tell who's talking, and immersion drops. Ridi and Millie reject such content as "below minimum quality standards." A web novel platform representative reported, "We submitted 30 audiobooks made with free TTS to Ridi—29 were rejected."

Another issue is unnatural emotional performance. Free TTS repeats a falling intonation pattern at the end of every sentence and pauses mechanically before commas and periods. Listeners describe it as "sounding like a robot reading" and request refunds. One study found that AI-voice-only audiobooks have refund rates over 3× higher than standard audiobooks. Platforms automatically filter or remove content with high refund rates, even post-launch.

International research on AI dubbing quality also points out that voice synthesis alone cannot meet the need for human expertise in synchronization and timing, voice consistency and expressiveness, narrative cohesion, sound design, and language and translation.

5 Quality Standards Ridi and Millie Demand—What Exactly Do They Check?

So what specific criteria do platforms use to judge audiobook quality? Here are the five standards Ridi and Millie commonly require:

1. Character Voice Differentiation

Main characters must be distinguished by at least 2–3 different voices. The male lead, female lead, and key supporting roles must each have distinct voices so listeners can tell who's speaking. For web novel audiobooks, multi-voice casting is virtually mandatory. Polita offers up to 14-voice multi-cast for web novel audiobook production, assigning voices that match each character's personality and age. One web novel platform reported that audiobooks produced with Polita's 14-voice casting were approved for registration on both Ridi and Millie and maintained listener ratings above 4.5.

2. Natural Emotional Performance

Emotional shifts must match the dialogue context. Joy, sadness, anger, surprise—emotions must come through naturally, and sentence-end intonation must not repeat mechanically. Polita fine-tunes emotional performance during the expert editing and QC stage after AI dubbing generation. The Pro plan, which recreates the timing, breathing, and emotion of the original voice actor, is also applicable to web novel audiobooks and delivers nuanced emotional expression for each character.

3. Appropriate Use of Background Music and Sound Effects

Background music and sound effects must be used appropriately to mark scene transitions and convey atmosphere. Free TTS outputs voice only, but platforms recommend audio elements to enhance the listening experience. Polita includes background music and sound effects as standard in audiobook production, selecting and inserting audio that fits the mood of each scene. This is a core element that boosts listener immersion and meets platform quality standards.

4. Terminology Consistency

Proper nouns, character names, and worldbuilding terms must be pronounced consistently throughout the audiobook. Free TTS often pronounces the same word differently each time. Polita provides glossary-based terminology consistency, managing web novel proper nouns and character names to be pronounced identically across the entire work. This is a core Polita capability also applied to game and educational content.

5. Voice Stability

Voice must remain stable without noise, dropouts, or volume fluctuations. Free TTS frequently suffers from mid-sentence cutoffs or sudden volume spikes. Polita checks voice stability during the expert QC stage after AI dubbing generation, performing noise removal and volume normalization. Up to two free revision rounds are provided to ensure quality fully meets platform standards.

Web Novel Audiobooks—How to Balance "Fast" and "Right"?

Web novel platform managers face a dilemma. They want to convert IP into audiobooks quickly for monetization, but free TTS gets rejected by platforms. Yet hiring professional voice actors for recording is too costly and time-consuming. The key to resolving this dilemma is the "AI dubbing + human polishing" approach.

Polita generates voice quickly with AI dubbing, then has experts edit and QC to meet platform quality standards. For web novel audiobook production, the Standard plan is $7–13 per thousand characters and the Pro plan is $17–23 per thousand characters, with a standard turnaround of approximately 5 business days (for 10-minute content). AI dubbing generation completes in 1–2 days, followed by expert editing and QC before delivery. There is no minimum order, so you can start with a single short web novel, and long series can be produced with consistent quality.

One web novel platform converted over 300 web novels into audiobooks through Polita. The representative stated, "We first tried free TTS, but Ridi kept rejecting our submissions. After switching to Polita, our registration approval rate hit 100%, and listener ratings averaged above 4.3." Polita uses glossary-based terminology consistency to manage character names and worldbuilding terms for identical pronunciation across the entire work, and multi-voice casting differentiates main characters to boost listener immersion.

Polita also provides an NDA-based secure workflow, safely handling unreleased web novel IP. Web novel platforms often launch new IP as audiobooks first, and Polita's experience handling sensitive content like pre-broadcast dramas and unreleased games ensures rigorous security management.

What Are Your Options for Producing Audiobooks That Meet Platform Quality Standards?

When producing web novel audiobooks, publishers have three main options:

First, hire professional voice actors for recording. Quality is highest, but costs range from tens of thousands to hundreds of thousands of won per minute, and recording, editing, and mixing take at least 2–3 weeks. This is difficult to choose unless you're a major publisher or working with bestseller IP.

Second, use free TTS sites. Costs are nearly zero, but as explained above, the output doesn't pass platform quality standards. Registration rejections delay launch, and rework costs end up causing losses.

Third, use AI dubbing + expert polishing. Services like Polita generate voice quickly with AI, then have experts edit and QC to meet platform standards. Costs are 30–50% of professional voice actor recording, turnaround is approximately 5 business days, and platform registration approval rates are high, enabling stable launch schedule management.

Polita supports over 70 languages and has processed over 45,000 video and audio content pieces annually. Beyond web novel audiobooks, Polita has production experience across animation, broadcast and OTT, games, education and e-learning, and brand films, giving it precise understanding of platform quality standards. Quote responses are provided within 1 business day, and free samples allow you to preview quality in advance.

The audiobook market continues to grow, but platform quality standards are rising alongside it. Publishers must choose partners who can produce "worth listening to, fast," not "fast and many." Rather than waste time with free TTS, producing to platform standards from the start is ultimately faster and more efficient. Polita combines the strengths of AI and human expertise to be a reliable partner for converting web novel IP into audiobooks.

FAQ

Q. Can I register an audiobook made with free TTS on Ridi or Millie?

Audiobooks produced with free TTS often fail to meet platform minimum standards for character differentiation, emotional performance, and background audio quality, resulting in registration rejections. Ridi and Millie have specified quality standards since 2025, rejecting submissions that fall short or removing content post-launch if refund rates are high.

Q. How much does web novel audiobook production cost?

By Polita's pricing, the Standard plan is $7–13 per thousand characters and the Pro plan is $17–23 per thousand characters. This is 30–50% of professional voice actor recording costs and includes multi-voice casting plus background music and sound effects. There is no minimum order, so you can start with a single short work.

Q. How long does audiobook production take?

Polita's standard turnaround is approximately 5 business days (for 10-minute content). AI dubbing generation completes in 1–2 days, followed by expert editing and QC before delivery. For urgent cases, the AI Only plan offers the shortest possible turnaround.

Q. Is multi-voice casting really necessary?

Web novel audiobooks often feature heavy dialogue and multiple characters, making multi-voice casting virtually essential. Single-voice production leaves listeners unable to tell who's speaking, reducing immersion and risking platform registration rejection. Polita offers up to 14-voice multi-cast.

Q. What's the platform registration approval rate for audiobooks produced by Polita?

One web novel platform produced over 300 audiobooks with Polita and registered them on Ridi and Millie, maintaining a 100% registration approval rate. Polita precisely understands platform quality standards and meets them through expert editing and QC.

References

- Apple Books Audiobooks Formatting Guidelines — Apple Books specifies platform review criteria including labeling requirements for AI narration audiobooks, quality review, and distribution blocks for repeated or bulk content submissions.

- Apple Books Digital Narration — Apple officially states that digital narration audiobooks may be rejected or not published if they fail to meet quality standards.

- Digital narration for audiobooks — According to Apple's official guide, AI narration audiobook production requires expert review by linguists, quality control staff, and audio engineers beyond voice synthesis.

- Quality in interlingual AI dubbing: Exploring machine and human … — International research demonstrates that AI dubbing quality requires human expertise in synchronization, timing, voice consistency, emotional expression, and sound design.