SecretOpedia reviews are designed to explain suitability and trade-offs, not to produce a score without context. We use direct testing when available, research from reliable sources, and clearly labeled limits.
Every review should answer two practical questions: who is likely to find the product useful, and who may be better served by another option.
Review labels
Each review receives a label that describes the evidence and depth behind it. The label helps readers separate direct testing from early or research-led analysis. It should be changed when later work materially expands or narrows the evidence.
Hands-On Review
The product was directly used, the key functions relevant to the conclusion were tested, and testing limits are disclosed. A short or restricted test may still qualify only when the label does not overstate the depth of use. The article should state the account level, device, or other condition that materially limited testing.
Research-Based Review
The assessment relies mainly on current official information, credible public evidence, independent analysis, and community reports. It does not imply direct use that did not occur. Conclusions must remain proportionate to the quality and consistency of the available evidence.
First Look
An early assessment based on limited access, an announcement, or newly available information. Conclusions are provisional and may change after broader evidence becomes available. A First Look should not receive the same confidence as a mature review when major functions remain untested.
Comparison
A structured evaluation of two or more services against the same practical needs. A comparison should not declare one product suitable for everyone; it should explain which reader profile each option may fit. Material differences in testing depth, plan level, or checking date must be disclosed.
Update Analysis
An assessment of a product change, new feature, policy shift, or interface change. It focuses on likely user impact and does not automatically replace a full review. Broader conclusions should wait until the change can be tested or supported by reliable evidence.
Evidence used in a review
Evidence may include official pages, help material, published terms, announcements, direct testing, screenshots created during testing, public user reports, videos, and independent reviews. We consider source date, directness, incentives, consistency, and whether a claim can be reproduced.
Official sources normally control for stated facts, while credible community evidence can reveal practical behavior or recurring problems. A user report remains a report unless separately confirmed.
Core scoring criteria
Where a numerical review is appropriate, SecretOpedia scores the product out of 10. A typical review considers:
| Criterion | Typical weight | What we examine |
|---|---|---|
| Features | 15% | Scope, usefulness, meaningful limitations, and whether the feature set supports the stated purpose. |
| Ease of use | 10% | Onboarding, navigation, clarity, accessibility, and the effort needed to complete common tasks. |
| Conversation experience | 15% | Coherence, repetition, character consistency, responsiveness, and practical control over interactions. |
| Image generation | 15% | Output consistency, control, visible defects, waiting experience, and usefulness for intended scenarios. |
| Customization | 10% | Available controls, flexibility, character creation, settings, and the cost of achieving a desired result. |
| Privacy information | 10% | Clarity of public privacy terms, account controls, data choices, and disclosed limitations. This is not a legal compliance audit. |
| Pricing and value | 10% | Current published cost, included functions, restrictions, renewal context, and value for defined user types. |
| Reliability | 10% | Availability during testing, errors, consistency, and credible reports of recurring technical problems. |
| Support information | 5% | Availability and clarity of help material, contact routes, and documented problem-resolution options. |
The typical weights total 100%. If a criterion is not relevant or cannot be assessed with adequate evidence, its weight may be redistributed among the remaining criteria. The review should disclose that limitation instead of assigning an invented score.
Calculating the overall score
Criterion scores are weighted and combined into an overall score out of 10, normally rounded to one decimal place. The written analysis controls the meaning of the number. A small numerical difference does not prove that one service is meaningfully better for every reader.
We do not add hidden points for an affiliate relationship, popularity, or access to a commercial partner. Any departure from the usual weighting should be explained in the review.
How to read the score
| Score | Editorial meaning |
|---|---|
| 9.0–10.0 | Exceptional performance for the defined use case, with limitations that do not materially outweigh the strengths for the intended audience. |
| 8.0–8.9 | Strong overall, with clear value for suitable users and identifiable trade-offs. |
| 7.0–7.9 | Good in several important areas but dependent on user priorities, budget, or tolerance for limitations. |
| 6.0–6.9 | Mixed. Useful for a narrower audience, with material disadvantages that require careful consideration. |
| Below 6.0 | Significant limitations, weak value, insufficient evidence, or problems that make the product difficult to recommend for the stated use case. |
A score is not permanent. Product changes, corrected evidence, pricing changes, and broader testing can justify an update.
Pricing, subscriptions, and changing terms
Pricing and plan information is checked against available official information when the review is prepared or updated. Taxes, promotions, regions, account status, and provider changes can affect what a reader sees.
Readers should verify the checkout page and current terms. Free access is not necessarily without limits. Canceling a subscription is not the same as deleting an account, and a refund policy is not the same as a guarantee that every request will be approved.
Privacy and safety assessment
We assess the clarity and accessibility of public privacy information, disclosed account choices, and observable controls. SecretOpedia is not conducting a formal legal audit, penetration test, or certification unless an article explicitly states otherwise and explains the method.
Adult functionality is assessed together with stated rules and safety boundaries. NSFW capability does not mean the absence of moderation or legal restrictions.
Commercial independence
SecretOpedia participates in the Secrets AI affiliate program. Commission does not alter the score, ranking, disadvantage, or verdict. A favorable rating cannot be purchased.
A company may identify factual inaccuracies before publication but has no approval right over the score, conclusion, or supported criticism. Sponsored material is labeled and is not presented as an independent hands-on review.
Corrections and reassessment
Material evidence supplied after publication can trigger a reassessment. We may change a criterion score or overall score when the underlying facts change, a testing error is found, or broader evidence materially alters the conclusion.
Significant changes should be explained under the Corrections and Updates Policy. Questions about a review can be sent to [email protected].