SEO/GEO follow-up item 3 from PR #1189. Review feedback recommended splitting this into an independent PR.
Value assessment
Value: medium. Queries such as "PawWork vs Codex app" have strong commercial intent and are also the type of structured comparison content that AI systems may cite directly. Standalone comparison pages are easier for search engines to index and easier for AI systems to cite than a table embedded in README.
Risk: competitor facts change quickly. We have already seen mistakes around desktop apps, Codex / Cowork naming, and local model support. Each conclusion row must link to official sources and include a last-updated marker.
Cost: medium to high, because this is new page work plus source-backed fact checking.
Goal
Add independently indexable and AI-citable comparison pages where each claim is backed by official sources and remains maintainable.
Scope
In scope, candidate pages:
/compare/pawwork-vs-codex-app//compare/pawwork-vs-claude-desktop//compare/open-source-desktop-ai-agent/
Each page should use a fixed structure: H1, TL;DR with 3-5 bullets, evidence table with official source links on every row, use-case fit, last updated, and a facts-checked-against list of official docs or release notes.
Out of scope:
- Marketing exaggeration. All competitor claims must cite official information.
- Multilingual versions. Those can be added later under the multilingual issue.
Relevant files or context
site/src/pages/ - new routes- README.md existing "How PawWork Compares" table, which can be reused only after every claim is checked against sources
- Official sources such as
openai.com/codex/ and Anthropic product pages / release notes
Verification
- Every comparison claim has a clickable official source link.
- Fact checking finds no known mistakes like the earlier desktop-app / local-model support drift.
astro build passes.- The new pages enter the sitemap and have correct heading structure.
Execution mode
Investigate and propose a plan first - post the page list, page fields, and fact-source check results, then wait for an explicit approved comment before writing.
Value assessment
Value: medium. Queries such as "PawWork vs Codex app" have strong commercial intent and are also the type of structured comparison content that AI systems may cite directly. Standalone comparison pages are easier for search engines to index and easier for AI systems to cite than a table embedded in README.
Risk: competitor facts change quickly. We have already seen mistakes around desktop apps, Codex / Cowork naming, and local model support. Each conclusion row must link to official sources and include a last-updated marker.
Cost: medium to high, because this is new page work plus source-backed fact checking.
Goal
Add independently indexable and AI-citable comparison pages where each claim is backed by official sources and remains maintainable.
Scope
In scope, candidate pages:
/compare/pawwork-vs-codex-app//compare/pawwork-vs-claude-desktop//compare/open-source-desktop-ai-agent/Each page should use a fixed structure: H1, TL;DR with 3-5 bullets, evidence table with official source links on every row, use-case fit, last updated, and a facts-checked-against list of official docs or release notes.
Out of scope:
Relevant files or context
site/src/pages/- new routesopenai.com/codex/and Anthropic product pages / release notesVerification
astro buildpasses.Execution mode
Investigate and propose a plan first - post the page list, page fields, and fact-source check results, then wait for an explicit approved comment before writing.