[chaitanyagiri]

What & why

Evening run, day 40 (6 Nov 2026, Friday). Two new competitor pages, one fix, and the rebuilt docs/.

Stacked on #826. This branch starts from #826's head (87d4f63). Merge order: #826, then this one. Or merge only this one: it carries #826's commit, and I then close #826 as landed.

• New: orchestrator-inc-vs-munder-difflin (1,010 words, 3 scenes and a title card). Orchestrator.inc, formerly Agent Orchestrator, is a desktop app (Apache 2.0) with phone apps. Its main agent plans the work and starts a worker for each piece in its own worktree. The page is a fair side by side, with prices dated 6 Nov 2026 (Free local, Starter $20, Pro $40, cloud only). It is the first of the competitor posts the founder asked for, one per product on the tracker, High threat first.
• New: emdash-alternatives (1,136 words, 3 scenes and a title card). Emdash renamed its desktop app Emdash Classic on 5 Nov 2026 and moved new development to Emdash 2.0 (early access). Six picks: Orca, Munder Difflin (second), Orchestrator.inc, T3 Code, Worktrunk, Claude Code --worktree.
• Fix: cursor-alternatives now dates the Windsurf rename to 2 June 2026, from Cognition's post "Windsurf is now Devin Desktop". It used to read as if the rename happened on 8 Oct.
• Links: one sentence in best-claude-code-multi-agent-tools to the Orchestrator.inc page, and one in worktrunk-alternatives to the Emdash page.

Type of change

• Bug fix
• New feature
• Refactor / cleanup
• Docs
• Build / CI

Checks

• gate.mjs: both new pages, cursor-alternatives and worktrunk-alternatives pass. best-claude-code-multi-agent-tools fails on length, description, question headings and pillar link. Those faults are from before this PR: the page already had 1,839 words on seo-engine: day 40 AM, daily brief for 6 Nov 2026, Conductor and Windsurf alternatives #826's head. This PR adds only one sentence, and the page is queued for a rewrite.
• Fact review: FAIL on 6 points, all fixed with the reviewer's own text. The fixes: the phone tunnel is on Starter and up only; a Munder Difflin line went beyond the README; a caption and a table intro implied Munder Difflin gives every agent its own worktree; Munder Difflin was third in one pick list; one caption had the wrong source.
• Scenes: each scene builder checked the frames at 25%, 50% and 75% and every still; I checked the rendered pages. Tap was not tested in a real browser.
• npm run build: passed.

Notes

• Two new pages, not three. With one subagent at a time, a third page would not fit in the 150 minute cap.
• Refresh queue: codex-pricing and grok-bot-pricing were skipped because help.openai.com and x.ai both return 403 (recorded, not worked around). is-claude-code-open-source needs Search Console data.
• OpenSEO is signed out, so there was no Search Console pull and no learn step.

Evidence

Images are on branch evidence/day-40-pm.

Before

The blog hub and cursor-alternatives before this change (#826's head).

After

The blog hub, both new pages, the fixed page, one scene from each page and both title cards.

🤖 Generated with Claude Code