How the current catalog is researched
For each tool, we review the vendor’s current product and pricing pages, describe the capabilities those pages support, and record the date checked. We then separate our best-fit and consideration judgments as editorial inference.
The current catalog is a focused six-tool sample, not a complete market survey. It currently includes one limited first-hand ChatGPT writing test. The exact prompt, conditions, three outputs, review, and limitations are published on the ChatGPT overview. The other five tools have not been tested directly.
What the evidence labels mean
- Official source
- A product, pricing, policy, or help page published by the vendor.
- Editorial inference
- Our interpretation of who a tool may suit, what tradeoffs matter, or how a buyer could evaluate it.
- Third-party observation
- Evidence from an independent publication, customer, analyst, or dataset. The current catalog does not use this label yet.
- First-hand test
- A task AI Stack Route performed directly with recorded inputs, conditions, outputs, human review, and limitations.
How we handle changing information
AI products and prices change quickly. Every evidence record includes a checked date, and the vendor source remains linked so readers can confirm important details before buying. A future monitoring system can flag stale records, but that automation is not active today.
When we call a product tested
A product can appear in the directory after an official-source review. We use “tested,” publish a performance finding, or make a stronger recommendation only after AI Stack Route completes and records a real task with the product.
The minimum record names the product plan and version when available, the test date, the task and inputs, the expected result, the observed result, the human review required, and material limitations. A single successful output does not support a broad quality claim. Comparative claims require the same task, criteria, and review process for each product.
Editorial independence
Commission size does not determine a product’s fit, ranking, or inclusion. Commercial terms will remain separate from product evidence. We do not publish ratings, testimonials, or performance claims without an evidence process that supports them.
See the separate affiliate disclosure for how compensation may work later.