Trust
Corrections log
When an independent review finds an error, or a reader reports one, we fix it and record it here with the date.
These entries come from independent audits in which a reviewer who did not write a page re-read each claim against the vendor's own page. To report an error yourself, use contact. Individual pages also carry their own update log.
- 29 September 2026: Price consistency across the site. We re-read changed prices on the vendors' own pages and confirmed Replit Core, Base44, v0, Claude Max and Codex figures. We removed a "5x limits" label that the Codex page does not carry, and updated three pages that said Softgen's pricing could not be read, because it can be read in a browser.
- 29 September 2026: Cursor pricing explained and AI website builder roundups. We sourced an attribution and added a caveat that v0's $90 price is a struck-through figure.
- 29 September 2026: Atoms review and Atoms vs Lovable. We removed or sourced three claims that our references did not support.
- 29 September 2026: Base44 vs Bolt.new. We fixed an internal contradiction and a stale date.
- 29 September 2026: Lovable pricing and authentication guides. We made the credit-expiry wording precise and added a method note to the authentication guide.
- 29 September 2026: What happens when AI builder credits run out. We removed two claims about Base44 (that a project stays live and that running out only blocks building) because the sources did not say so, and replaced "not stated" for Atoms rollover and top-ups with what its pricing FAQ says.
- 29 September 2026: Best AI app builder and best AI code editor roundups. We removed three unsupported claims from the app-builder roundup and corrected a stale detail in the code-editor roundup.
- 29 September 2026: Claude Code vs Cursor. We corrected a statement about using Claude Code inside Cursor that Anthropic's documentation contradicts.
- 29 September 2026: Backend tools and bridges. We removed an overclaim about hosting and re-read the Supabase, Neon, Stripe and Hostinger pages.
- 29 September 2026: Bolt, v0, GitHub Copilot and Base44 pricing guides. We corrected Bolt's free custom domain from "not listed" to "no", removed an unsourced "no team features" line on v0, and updated GitHub Copilot's business plan figures now that the June to August 2026 promotion has ended. On Base44 we replaced one custom-domain claim with the conflict between the pricing page and the blog, and stated that the backend-function plan requirement also conflicts between sources.
- 29 September 2026: Windsurf review. We removed a roughly 17 percent annual saving that no Windsurf page states, replaced SWE-1.5 with SWE-2 (the only model the pricing page names), and corrected Teams from $40 per user to an $80 base plus $40 per seat.
- 29 September 2026: Codex vs Claude Code. Claude Code was described as starting in manual mode; its documentation makes auto mode the starting mode from version 2.1.283. We rewrote the summary, table and the safety pick, and listed all three published Codex allowance tiers.
- 29 September 2026: Cline pricing explained. We corrected a claim about seat fees (Cline states none only for Open Source), replaced a wrong comparison between a heavy day and a $20 subscription, and fixed a daily-cost comparison that did not match the model rates.
- 29 September 2026: Claude Code, Cursor rules and CLAUDE.md guides. We corrected the earlier default permission mode for Pro, Max and Team plans, and corrected how Claude Code reads AGENTS.md: only when there is no CLAUDE.md, unless a setting or an import says otherwise.
Pages that were re-read and needed no change, such as the Claude Code pricing, Replit pricing, Lovable vs Bolt vs v0 and AI builder lock-in guides, are not listed.