About PodcastFactor
PodcastFactor is an editorial site that helps creators choose AI podcast tools. This page explains how our reviews are made today, how we are testing tools, and how we make money.
How Our Reviews Are Made Today
PodcastFactor's existing reviews were written from vendor documentation, public information and limited use. We are re-testing each tool with the published protocol below; every review will say whether it has been tested, when, and on which plan. We earn affiliate commission from some vendors listed on this site; that never changes rankings.
Tested under our protocol
None yet.
Last full price check: October 5, 2026.
Our Editorial Rules
Editorial Independence
Affiliate relationships never change a ranking or a score. Rankings sort by score within a tool category, and the list of vendors that pay us commission is published below.
Fact, Test Result and Opinion Kept Apart
Every page separates verified facts (from vendor documentation, dated), logged test results, and the editor's opinion. Vendor-reported benchmarks are labelled as vendor claims.
Published Testing Protocol
Every tool in a category gets the same inputs so results are comparable and reproducible. A review carries a tested score only when a test log backs it.
Dated Prices
Every price shows the date we checked it on the vendor's own pricing page. We aim to re-check prices on every tool page monthly.
How We Test
This is the protocol we are now running on every tool. Every tool in a category gets the same inputs, so results are comparable and reproducible. We store the inputs and outputs and publish samples.
| Category | Shared input | What we measure |
|---|---|---|
| Generators | One 2,000-word article and one 900-word two-host script | Factual errors against the source (count), script control, voice naturalness (blind panel of 5+, rated 1–5), retakes needed, minutes from start to export, credits used, cost per finished minute |
| Voice engines | One 600-word narration with names, numbers, a question and an emotional turn | Mispronunciations (count), pacing, emotion, artifacts, tag adherence, cost per finished minute |
| Editors | One 20-minute two-person recording with fillers, crosstalk, level mismatch and room noise | Fillers caught and missed, false cuts, cleanup quality (blind A/B), time to finished edit, export formats |
| Recorders | One remote session with the guest on weak Wi-Fi | Local vs cloud track quality, dropouts, sync, guest setup steps |
Each test log records the tool, plan, version or model, date, tester, settings, input file, output file, measurements, issues, time spent and cost. A review cannot carry a tested score without a linked log.
Scoring Rubric
Once a tool is tested, its score is a weighted sum of logged measurements. Weights differ by category because the job differs, and tools in different categories are never ranked against each other. Scores on untested pages are earlier editorial estimates and are labelled as such.
| Criterion | Generators | Voice engines | Editors |
|---|---|---|---|
| Output quality (blind panel, error counts) | 35% | 45% | 35% |
| Control and editing | 20% | 15% | 25% |
| Workflow speed | 15% | 10% | 20% |
| Price per finished minute | 20% | 20% | 10% |
| Licensing and policy clarity | 10% | 10% | 10% |
How We Make Money
PodcastFactor is funded through affiliate partnerships. When you click certain links on our site and make a purchase, we may earn a commission at no additional cost to you. Whether a vendor pays commission never affects its score or its position in a ranking.
We are compiling the list of vendors that pay us commission and will publish it here.
We also have a commercial interest in these products, which we review under the same rules: they are not ranked, and they get no score until a logged test backs one.
- Sintra Studio (sintrastudio.com)
Contact Us
Found an outdated price or a factual error? Have a tool you would like us to test? We would like to hear from you.