How this guide is built
Evidence before interpretation
Our source hierarchy, claim labels, review cadence, conflict rules, and limits for a product that can change faster than a static page.
Every material claim should answer four questions: who said it, where is the closest source, when was it checked, and what qualification keeps it from implying more than the evidence supports?
Source hierarchy
- Official product records.xAI release notes, product pages, legal pages, and provider pricing are primary for launch status, named features, and listed access.
- Independent reporting or measurement.Dated reporting is used to corroborate the record, surface omissions, and identify open evaluation questions.
- Practitioner and community discussion.Useful for terminology, concerns, and FAQ discovery; never proof of a controlled product fact.
- Editorial interpretation.Our synthesis is explicitly framed as guidance or a limitation, not as a provider claim.
Claim labels
“Official” means the relevant provider published it. It does not mean the claim was independently tested. “Independent” means a third party reported or measured something with enough method to inspect. “Community” means an individual experience or discussion shaped the questions, not the answer.
Claim: xAI says Bots can continue working after the user's laptop is closed.
Qualification: the reviewed launch record does not publish a completion-rate measurement for that behavior.
Review cycle
Prices, plan names, platform availability, beta status, and documentation links are time-sensitive. Each page displays a review date. A later change should update the claim, qualification, structured data, and source ledger together.
When official pages conflict, the newest page closest to the claim wins. If the conflict remains unresolved, both versions are described rather than silently merged.
Corrections and limitations
This first edition is based on public material available immediately after launch. We do not have privileged access, private documentation, a provider relationship, or a standing independent benchmark suite. That limitation is why product behavior is attributed and why the first-workflow checklist is conservative.
Editorial changes should be traceable in the site's Git history. Material corrections update the relevant page and its review date; silent factual rewrites are not the goal.