Editorial standards
Evidence before enthusiasm.
Aivunex helps readers choose AI tools and SaaS products with clear evidence, practical trade-offs, and transparent limits. This page explains how we research, evaluate, score, update, and disclose our work.
Our goal is not to identify one universal “best” product. A tool can be excellent for a solo creator and a poor fit for a regulated team. We define the intended user and job first, then judge how well each product serves that situation.
What our research labels mean
Used in a representative workflow
We accessed the product and completed relevant tasks. We explain what was tested and avoid implying that a short evaluation represents every feature or long-term use case.
Verified without pretending we tested it
When hands-on access is unavailable or inappropriate, we use official documentation, pricing pages, release notes, policies, and clearly identified supporting sources.
Time-sensitive claims were rechecked
Pricing, plan limits, availability, system requirements, and important product claims are checked on the date displayed in the article.
Our evaluation framework
When Aivunex assigns a score, it uses five weighted areas. The written verdict and fit for a specific reader matter more than a small difference between two numerical scores.
How a review is produced
- Define the reader and task. We establish who the product is for, what problem it should solve, and which alternatives genuinely compete with it.
- Check primary sources. We review the vendor’s current documentation, pricing, release information, policies, and technical requirements.
- Test a realistic workflow when access is available. We focus on setup, core output, friction, limitations, and the controls that matter for the stated use case.
- Record trade-offs. Reviews must explain meaningful weaknesses, ongoing costs, privacy implications, maintenance requirements, and situations where another tool is better.
- Compare like with like. Products are evaluated against tools serving the same job—not against unrelated software with a different audience.
- Verify and update. We display an update or verification date and revise material claims when products, pricing, or availability change.
How to read our scores
| Score | Meaning |
|---|---|
| 4.5–5.0 | Outstanding for the stated user and use case, with few important compromises. |
| 4.0–4.4 | Strong recommendation, with limitations readers should understand. |
| 3.5–3.9 | Good for a narrower audience or budget, but alternatives may fit many readers better. |
| 3.0–3.4 | Useful in specific circumstances, with material trade-offs. |
| Below 3.0 | Limited recommendation until important weaknesses are resolved. |
Pricing and availability
We prefer official pricing and documentation. “Free” is separated from free trials, limited credits, open-source software, and self-hosting that still requires hardware or operational work.
Affiliate independence
Affiliate relationships do not determine scores, rankings, inclusion, or conclusions. Potential commissions are disclosed and do not change the price a reader pays. Read our affiliate disclosure.
Corrections and updates
If you find a factual error or important change, contact the Aivunex editorial team. We review evidence and correct material inaccuracies rather than silently preserving an outdated claim.
See the framework in practice
Start with Aivunex’s current AI-agent research and compare the evidence, costs, and trade-offs.