How to use Claude to prototype a search intelligence tool
The author prototyped search intelligence with Claude and Claude Code, encountered excessive API consumption, then used SerpAPI and Lovable for a tool used daily.
Added1125 use casesShowing 313–336
Preparing search…
The author prototyped search intelligence with Claude and Claude Code, encountered excessive API consumption, then used SerpAPI and Lovable for a tool used daily.
AddedIn a reported demo, AURA investigated checkout 502s, traced them to a downstream memory leak, documented a fix through a human-gated GitHub call, and checked recovery.
Added AURAOutscal constrained agent tools and supplied context upfront after inconsistent Claude Code runs, then moved its React video pipeline to the Claude Agent SDK and launched an automated web version.
AddedRecall built Predict largely with Claude Code subagents, then shifted from a Figma MCP design workflow to a designer prompting iterations directly into code. The author reported an end-to-end working loop with rough edges.
AddedHawkDoc combines Lexical editing with React, collaborative-sync infrastructure, and Markdown, HTML, and PDF export. Its author credits Claude Code for early editor integrations and export development.
AddedA developer used Claude Code and Codex CLI to write svelte-check-rs, retaining architecture and real-codebase testing. The source reports cold, warm, and iterative speed improvements on a SvelteKit monorepo.
AddedGodogen combines two Claude Code skills with a custom Godot reference system, headless scene construction, and separate Gemini Flash screenshot evaluation to generate playable Godot 4 projects.
AddedIndexio indexes a codebase and serves Claude Code lookups through MCP, falling back to regex when needed. Its authors report reduced token use but identify symbol-heavy workflows as a limitation.
AddedChatML gives parallel Claude Code tasks separate Git worktrees, branches, and working directories. The team used the workflow for most of ChatML and reports merging over 750 generated pull requests.
AddedMaxxwell coordinates parallel Claude and Codex sessions through an orchestrator that monitors goal checklists and brings important decisions back to the user.
AddedWillison used Claude Code to narrow a Cloudflare challenge to multi-parameter searches. When the MCP interface could not edit the relevant rules, the agent switched to the API.
AddedWillison asked Claude Code for web to compile Gifsicle to WebAssembly and build a preview interface. The agent used Rodney to exercise the page and fix UI bugs during implementation.
AddedWillison wanted to optimize a phone recording before publishing it. He used Claude Code for web to build a browser tool around FFmpeg compiled to WebAssembly, with multiple output presets.
AddedWillison delegated a Moebius model port to Claude Code using local repositories and a research note. He tested the browser UI and returned errors while the agent iterated on ONNX and WebGPU support.
AddedWillison needed a video of an interactive canvas animation for a keynote slide. A local Claude Code session used Playwright to open the page, time clicks and produce a recording.
AddedAnthropic and Andon Labs let a Claude-based agent manage a real office shop. It searched for products, set prices and requested restocking, but the researchers report losses and persistent mistakes.
AddedIn the next office-shop phase, Claudius gained browser access, CRM and better inventory context. It checked supplier prices and delivery details while human purchase approval remained mandatory.
AddedSimon Willison gave Codex a specific defect in his Django blog: social preview descriptions exposed raw Markdown. Codex changed templates and added a passing test in a configured PostgreSQL environment.
AddedWillison published the Codex session used to build the Datasette project blog. The agent inspected the existing site, moved conflicting database routes and implemented blog behavior with local checks.
AddedWhen crawlers hammered datasette.io, Willison used Codex to build a configurable IP rate-limiting plugin. His release note includes the configuration used on the live site.
AddedWillison supplied an existing parser and an independent compatibility suite, reviewed an API plan, then let Codex implement a JavaScript port through repeated tests and commits.
AddedDodds used Cursor with GPT-5.4 to add lexical retrieval to his semantic search. Production blocking exposed an architectural flaw, so later agent work moved search into a unified Worker.
AddedDodds delegated an audio pipeline migration to Cursor. He rejected an initial design that preserved a separate service and used another agent run to make the sandbox part of the existing queue Worker.
AddedWillison ran o4-mini-deep-research to find surviving orchestrions and return structured location records. The model used web search and code execution; he explicitly considered the resulting list incomplete and unverified.
Added