NaturalReader

The plan table on the product page lists Free at $0 with unlimited listening in the free voices, 5 minutes a day of the better Plus AI voices, limited AI features and annotations, with no card required. The features article adds that extended listening, MP3 conversion, OCR for scanned files and full Text Filter controls need a subscription.

The same table ticks Voice Cloning in the Free column, while the pricing article lists cloning only on Plus and Pro and says Lite excludes it. Treat cloning on a free account as unconfirmed until the app offers it. Use Free to check that your files open and that a voice suits you before paying.

Every paid tier includes MP3 conversion, but each voice tier has its own pool. The tiers differ mainly in which voices can be exported and whether cloning is included:

• Lite: unlimited listening and 1 million MP3 characters a month, both with Lite voices only; no cloning.
• Plus: adds Plus voices and up to 2 cloned voices, with 500,000 characters a day of listening and 1 million MP3 characters a month shared between them.
• Pro: adds the newest HD Pro voices and Reading Styles, which set tone, emotion and accent by prompt; the same shared pools apply.

Editorial assessment

Choose Lite if Lite voices sound acceptable for your material, since it already gives a full monthly MP3 pool. Pay for Plus only when you need a clone or a noticeably better voice, and Pro when Reading Styles matter.

Not with a personal plan. The subscription agreement says audio from the Personal version may be used only by the authorized user for private listening, and may not be published, embedded, sold or shared with another person. It also excludes videos, podcasts, e-learning, presentations, websites and phone systems, and says even internal or nonprofit redistribution needs a Commercial license.

The checkout page repeats that MP3 files cannot be used for non-personal purposes. EDU audio follows the same rule: a school needs a Commercial license before putting generated audio into lessons, videos or student-facing media.

Plus and Pro can clone up to 2 voices from an upload or recording, and the features article limits cloned voices to personal use. The agreement allows cloning only where you have valid legal rights and authorization, and forbids imitating a real person without informed authorization, false endorsements and impersonation.

Clone only your own voice, or that of someone who has agreed in writing to this use. The AI use policy says NaturalSoft may ask for evidence of authorization and may suspend or delete unauthorized voices, so keep that permission on file.

The AI use policy says customer text, documents, voice samples, cloned voices and generated audio are not used to train NaturalSoft or third-party models. Features may still be processed by providers named in that policy, including Google Cloud, Microsoft Azure Speech, ElevenLabs, OpenAI and Amazon Polly.

Retention figures differ between documents. The features article says your library keeps uploads for up to 1 year and audio for up to 1 month, while the privacy policy lists personal documents as kept for up to two years unless deleted. Delete confidential files yourself when you finish with them rather than relying on either period.

For personal plans bought on the website, you can ask for a full refund within 3 days of purchase or renewal on monthly billing, or within 7 days on yearly billing. After that there are no prorated refunds. Refunds are not automatic: request one through Account, then Contact Us, or by email.

Plans bought through the App Store or Google Play follow Apple’s or Google’s refund process instead. Subscriptions renew automatically, and after cancelling you keep the plan’s features until the end of the billing period, so cancel before the renewal date if you only needed one cycle.

1. Note the MP3 characters remaining before starting.
2. Upload the scan, run OCR and fix any misread words with the pronunciation editor.
3. Turn on the Text Filter for page numbers, headers and footnotes.
4. Convert the chapter to MP3 with one Lite voice and download it.
5. Play the file outside the app and record the characters used against the 1 million monthly pool.

Accept an MP3 that reads the chapter in order without page numbers or headers and whose character count matches the text length. Reject skipped columns, read-aloud headers or a file that will not play offline. No audio result is reported until the exercise is run.

For free downloadable speech with a published daily allowance, compare TTSMaker. For voiceovers you are allowed to publish, see Murf or ElevenLabs.