Practical Lessons from Building an AI Site Cloner
Unique-Watercress225 · reddit · 2026-07-15
The author shares key takeaways from building a **website cloning tool**, with a core thesis: once an AI agent hits 90% completion, the final 10% is often the hardest part. Main conclusions include: - **Don't rely too heavily on visual inference from screenshots**; models will guess spacing and colors, leading to massive discrepancies. - A better approach is to directly feed the model the **browser's final rendered structural data**, rather than letting it "look at an image and imagine" the layout. - **Fonts, SVGs, and images** will silently alter the layout, so you must fetch the real assets. - For the workflow, it's recommended to first clone into **pure HTML**, and then migrate to frameworks like Next.js or Vue. Finally, the author briefly introduces their tool, **Copy-Anything**: input a URL to generate a web clone with nearly 99% visual fidelity, which is open-source and free.
More from coding & agent
- Codex turns out 123 screensavers in one playful batch — intellectronica · 2026-07-21
- Grok Build adds `grok doctor`, resumable sessions and remote image paste — mark_k · 2026-07-21
- Autoresearch proposes packaging ML runs as studies with questions, analysis, and code diffs — morgymcg · 2026-07-21
- CHAP defines approvals, handoffs, and audit logs for human-agent workflows — DeliveryTechnical199 · 2026-07-21
- The author says Codex reached 20x and is now debugging spec decoding on a hybrid parallel setup — TheZachMueller · 2026-07-21
- Axcess adds an MCP connector for WCAG accessibility checks that scanners miss — modelcontextprotocol · 2026-07-21