Source Verification: Read source-0.html.gz (bun.com/blog/bun-v1.4.4, HTTP 200, snapshot sha256 e9af8ca2...) via raw HTML text extraction, plus a live re-fetch with curl (page bytes differ from the snapshot, as expected for a dynamic page; every key string re-confirmed live: '286x faster', '3x to 6x faster than oxfmt 0.74.0', '234x to 425x faster than Prettier 3.9.9', '9,731', '1,343', 'exits with code 2'). Verified individually against the post: byline Jarred Sumner and date October 22, 2026; verbatim quote 'bun fmt is an incredibly fast, Prettier-compatible code formatter that works with your existing Prettier or oxfmt config'; language list (JavaScript, TypeScript, JSX, Flow, CSS, SCSS, Less, HTML, Vue, Svelte, Angular, Lightning Web Components, MJML, JSON, JSONC, JSON5, YAML, TOML, Markdown, MDX, GraphQL, Handlebars) and embedded code; Prettier's own repo 2,616 files with their config, bun fmt --check 59.7 ms / 48 MB vs Prettier 3.9.10 17.1 s / 459 MB, 'That's 286x faster' (the post's prose rounds to 60 ms and 17 seconds; the table gives the article's precise figures); vitest, date-fns, monkeytype, h3 '234x to 425x faster than Prettier 3.9.9' (verbatim); oxfmt '3x to 6x faster than oxfmt 0.74.0' (verbatim), table RSSHub/redocly-cli/vite/fumadocs with fumadocs 31x; one-core table 1.7x (redocly-cli) to 21x (fumadocs); hardware 16-core M4 Max MacBook Pro for vite (mean) and Intel Xeon Platinum 8488C for the rest (median); config files read (.prettierrc*, prettier.config.*, 'prettier' key in package.json, .editorconfig, .prettierignore, .oxfmtrc.json/.jsonc, oxfmt.config.ts/.mts); --check, --staged, --changed; built-in import sorting; unsupported plugin such as prettier-plugin-jsdoc prints an error naming the plugin, skips affected files, exits code 2; lineage quote 'It's a port of Prettier, and the JavaScript and TypeScript printer is a port of oxc_formatter, which is a port of Biome's formatter' (verbatim) and thanks to the Prettier, Biome and Oxc teams; compatibility: CI runs Prettier 3.9.9's own formatting tests on every commit, 9,731 tests / 9,648 identical / 99.1% (defined as byte-for-byte identical output), oxfmt 0.72 on 1,343 of 1,498 fixtures (89.7%); intentional difference: Prettier 3.9.9 drops <!DOCTYPE html> in Handlebars, bun fmt keeps it; bun check '1.4x to 7.8x faster on large projects' (verbatim), MUI docs 10.28 s to 1.31 s, peak memory 11.2 GB to 1.3 GB; bun install converts Yarn 2+ (Berry) yarn.lock to bun.lock; retries up to five times, 50 ms first delay, doubling. All confirmed. The 'built into Bun' and 'Prettier-compatible' wording follows the post exactly (a bun fmt command; not described as opt-in or gated). The post's 330-repository comparison is not used by the article, so nothing depends on it. Read source-1.html.gz (GitHub release page, HTTP 200): title 'Bun v1.4.4', tag bun-v1.4.4; cited only to show the release exists, which it does. GitHub API: published_at 2026-10-22T10:11:32Z, not draft, not prerelease, confirming October 22, 2026. Internal link /article/2026-10/10-bun-143-adds-bun-check-... resolves to the published Bun 1.4.3 article and is described accurately (bun check type checker). suspicious_patterns: null for both sources. Allowlist: github.com (line 1131) and bun.com (line 2387, already listed with a first-party justification comment) both match by exact hostname; no allowlist change needed. Orphan-link diff: every URL in the body (bun.com blog, GitHub release) is in article.sources; only other link is the internal Herald path. Single-vendor sourcing for all performance and compatibility claims: no independent coverage cited; the article says so explicitly. Independent-source count not applicable (Briefing).