How We Research and What We Have Not Tested

NarratedIndie has not run any original tests of narration or writing tools as of 2026-11-08. Every product and policy page is based on published documentation, pricing, policies and reports, read on a stated date. No test results exist yet.

How research works now

1. Primary sources first. Pricing pages, help pages, terms and official reports from the vendor or store come before anything else.
2. Read dates. Each fact is tied to the date the page was read, and the page's own date is noted when it shows one.
3. Mark what could not be read. If a page was blocked, returned an error, or was only seen as a search snippet, the article says so and does not state its contents.
4. Never state what was not read. Silence in a source is reported as silence.
5. Secondary claims stay secondary. Blogs and vendor marketing do not establish rules or prices. Marketing claims are attributed to the vendor.

The wider rules are in the editorial policy, and how AI is involved is in the AI use policy.

What this means for you

Pages tell you what a tool's own documents say, not how its output sounds or how much work a finished book needs. For example, the AI audiobook cost calculator is built to work from published prices and your inputs, not from measured results.

Planned test protocol (a plan, not results)

The following is planned and has not been carried out. No results exist.

• One fixed text set. A fixed set of chapters from US public-domain works on Project Gutenberg, the same for every tool.
• Same settings per tool. Settings are recorded and held as close as each tool allows.
• Recorded measures. Credits used, number of regenerations, edit hours per finished hour of audio, and export specifications.
• Dated and labelled. Each result will carry its test date and the plan tier used.

When results exist, they will be published with that detail and the relevant pages will be updated and logged in the changelog. Until then, no page claims a measured result.

Return to the home page.