Borderline tool polish — moat-strengthening for the keepers

Cull-pass on 2026-05-15 removed 13 tools (7 hard-cull, 6 soft-archive). resize-image + resize-all-platforms were absorbed into bulk-image-resize with 301 redirects. A handful of tools survived the cull because their moat is real — but the moat needs polishing for the survival to hold up post-launch. This phase tracks that work.

The pattern: every tool listed below “looks commodity at first glance” — exactly the kind of tool a competitor might claim AI replaces. The fix isn’t to cull; it’s to lean harder into the specific niche that makes the tool durable.

• Audio editors (audio-speed-changer, audio-pitch-changer) — VLC and DAWs do these, but our moat is no install + privacy for podcast / language-learning / music-practice users. Need export presets that match the actual workflows (1.5x podcast, semitone music intervals).
• uuid-generator — uuidgen exists. Our moat is bulk + copy-list for devs who’d rather not write a one-liner. Make the bulk path one click.
• hash-generator — openssl dgst exists. Our moat is privacy (no upload). Surface that explicitly + accept file inputs alongside text.
• qr-code-generator — every CDN-script QR tool exists. Our moat is logo-embed with proper error correction. Verify it works at common sizes.
• altitude-baking-calculator + canning-calculator — niche audiences, but the stakes (chemistry / safety) are real. Need data quality + UX that respects the audience (elevation lookup, USDA-source canning tables).
• photo-collage-maker — Canva owns this workflow. Needs a real template library to compete; if the library doesn’t ship, this is a cull candidate next round.
• compare-pdf — featured tool, strong moat, but missing the annotated-diff-PDF export that legal teams ask for first.

Cull-revisit trigger: if these aren’t polished within 2 release cycles after launch, re-evaluate against the D18 moat framework. Survivors that don’t earn their keep get soft-archived in the next sweep.