Grok Bot flagged two fare drops. @tomkarren repriced the flights himself and reports about $400 saved on a trip for two. @xai agentsinuse.com/use-cases/gr…

Toll charges become an emailed reimbursement PDF. @coyotespike describes OpenClaw preparing the attachment, with the person still responsible for attaching it to a claim. @openclaw agentsinuse.com/use-cases/op…

15 pages read, 3 tools shortlisted. @Jtyles reports Hermes Agent saved the research workflow as a weekly skill and shared its instructions. Later runs were not observed. @NousResearch agentsinuse.com/use-cases/he…

Family meal planning starts with a weekly survey. @RobertJCrane reports Grok Bot collates preferences, checks the calendar and returns a meal plan and grocery list. @xai agentsinuse.com/use-cases/gr…

Claude Code helped coordinate invoice chasing. The author reports about ten minutes of work for 50 monthly invoices, versus five hours manually. Draft emails still needed approval. @claudeai agentsinuse.com/use-cases/cl…

34 prospective tenants contacted, 8 tours scheduled. @AlexPorrett reports Grok Bot handled rental enquiries and tour coordination, not a completed tenancy. @xai agentsinuse.com/use-cases/gr…

Grok Bot flagged a duplicate landscaper bill. @multifamjunk reports a credit just under $700 after the bookkeeper followed up. Most flags were legitimate. @xai agentsinuse.com/use-cases/gr…

A photo, product dimensions and three or four test prints led to a custom doorbell cover with Claude. The maker reports a finished fit, not proof that it solved overheating. @claudeai agentsinuse.com/use-cases/cl…

Over 200 integration tests fixed in one week. A developer and coworker used Claude to match old-system behavior, then reviewed the code again before launch. @claudeai agentsinuse.com/use-cases/cl…

Roughly 100 failing mail-server conformance tests became a passing suite after a repair request to Claude, the author reports. About 45 minutes, no further prompts. @claudeai agentsinuse.com/use-cases/cl…

A wearable badge finally booted after its maker removed a shorting switch and corrected the build environment. Claude Code helped with firmware and logs, but missed hardware errors. @claudeai agentsinuse.com/use-cases/cl…

Seven revisions to a snap-fit enclosure. Claude worked from drawings and fit photos to correct vents, solder clearances and board retention. @claudeai agentsinuse.com/use-cases/cl…

Claude workers implemented workspace and handoff features from a plan, with a separate merger handling integration. The project was still early. @claudeai agentsinuse.com/use-cases/cl…

Claude Code helped investigate two SSRF flaws. The researcher verified the URL-boundary failures and worked with maintainers on fixes and private advisories. @claudeai agentsinuse.com/use-cases/cl…

Retailer screenshots plus a personal budget became a speaker comparison and PDF recommendation with Claude Code. The published result is a report, not a completed purchase. @claudeai agentsinuse.com/use-cases/cl…

Multi-monitor support and a system tray for a custom window manager. Its author reports daily use of the Claude-written additions, with cleanup pending. @claudeai agentsinuse.com/use-cases/cl…

Playable game projects from text prompts. One developer built a Claude Code pipeline with engine-specific references and separate screenshot checks for visual faults. @claudeai agentsinuse.com/use-cases/cl…

Claude wrote most of a portable terminal workspace with a one-command setup. The author set the rules and made the decisions. @claudeai agentsinuse.com/use-cases/cl…

30 financial panels became 516 with Claude Code. The important caveat is that most added panels use simulated or seeded data, not live feeds. @claudeai agentsinuse.com/use-cases/cl…

Testing Claude Code configurations exposed misleading pass reasons, grader errors and an actual container teardown from an evaluation workspace. @claudeai agentsinuse.com/use-cases/cl…