<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Agents in the Wild</title><link>https://agentsinthewildai.com/</link><description>Real things people get AI agents to do, with the source and a prompt to try it.</description><language>en</language><item><title>Meta Muse hunts apartments daily, only pings when it's worth it</title><link>https://agentsinthewildai.com/cases/288-meta-muse-hunts-apartments-daily-only-pings-when-it-s-worth/</link><guid>https://agentsinthewildai.com/cases/288-meta-muse-hunts-apartments-daily-only-pings-when-it-s-worth/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>Apartment hunting with a long list of requirements, they gave Muse the list and had it search Craigslist and Facebook Marketplace daily for places that fit. It only pings them when something worthwhile turns up.</description></item><item><title>A date string tool built with Gemini should save 45 seconds a week</title><link>https://agentsinthewildai.com/cases/287-a-date-string-tool-built-with-gemini-should-save-45-seconds/</link><guid>https://agentsinthewildai.com/cases/287-a-date-string-tool-built-with-gemini-should-save-45-seconds/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>Tired of typing new Date().toISOString() and doing timezone math by hand, they built a small tool with Gemini, which added obscure date formats and random historical events along the way.</description></item><item><title>One prompt to Claude, and 15 minutes later: fix it today or Tuesday?</title><link>https://agentsinthewildai.com/cases/286-one-prompt-to-claude-and-15-minutes-later-fix-it-today-or-tu/</link><guid>https://agentsinthewildai.com/cases/286-one-prompt-to-claude-and-15-minutes-later-fix-it-today-or-tu/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>In a meeting about a bug in an app they didn't own, they told Claude to find out if it was user error, notify the user, and tell the developer what to fix. Fifteen minutes later, the developer asked: today or Tuesday?</description></item><item><title>From $2-4 a day to under $0.25: CI moved to a homelab with Claude</title><link>https://agentsinthewildai.com/cases/285-from-2-4-a-day-to-under-0-25-ci-moved-to-a-homelab-with-clau/</link><guid>https://agentsinthewildai.com/cases/285-from-2-4-a-day-to-under-0-25-ci-moved-to-a-homelab-with-clau/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>After $45 in GitHub Actions overage in one month, a solo founder put an idle homelab to work as CI runners, with Claude making it easy to set up. Daily spend dropped from $2-4 a day to under $0.25.</description></item><item><title>15k pages of discovery: Claude Code finds the evidence, with a trail</title><link>https://agentsinthewildai.com/cases/284-15k-pages-of-discovery-claude-code-finds-the-evidence-with-a/</link><guid>https://agentsinthewildai.com/cases/284-15k-pages-of-discovery-claude-code-finds-the-evidence-with-a/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>About 15k pages of discovery stood between a solo practice and a summary judgment motion. Claude ran OCR, built embeddings and search scripts, then found evidence for each assertion, with traces to check.</description></item><item><title>Synth firmware came much faster with gentle prodding at Claude</title><link>https://agentsinthewildai.com/cases/283-synth-firmware-came-much-faster-with-gentle-prodding-at-clau/</link><guid>https://agentsinthewildai.com/cases/283-synth-firmware-came-much-faster-with-gentle-prodding-at-clau/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>A first homemade synth has 36 hall-effect keys and buttons and 7 rotary encoders. The firmware started by hand, then gentle prodding at Claude got everything needed much faster. Rev 2 of the PCB is nearly done.</description></item><item><title>Meta Muse installed duckdb and now runs their budget exports</title><link>https://agentsinthewildai.com/cases/282-meta-muse-installed-duckdb-and-now-runs-their-budget-exports/</link><guid>https://agentsinthewildai.com/cases/282-meta-muse-installed-duckdb-and-now-runs-their-budget-exports/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>Muse comes with its own VM, so they had it install duckdb and set up automated exports for their budgeting, which now run for them.</description></item><item><title>A six-page ChatGPT prompt does the first pass on businesses for sale</title><link>https://agentsinthewildai.com/cases/281-a-six-page-chatgpt-prompt-does-the-first-pass-on-businesses/</link><guid>https://agentsinthewildai.com/cases/281-a-six-page-chatgpt-prompt-does-the-first-pass-on-businesses/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>Buying small e-commerce brands, they codified everything they look for into a six page prompt. Given a listing's screenshots and financials, ChatGPT takes a first pass to decide whether a seller call is worth taking.</description></item><item><title>Claude invents new Fermi questions from random dictionary words</title><link>https://agentsinthewildai.com/cases/280-claude-invents-new-fermi-questions-from-random-dictionary-wo/</link><guid>https://agentsinthewildai.com/cases/280-claude-invents-new-fermi-questions-from-random-dictionary-wo/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>A browser game where you answer with a power of ten leans on Claude for code and content: fact checking answers, writing explanations, and inventing new questions.</description></item><item><title>Claude coded it: a 60-hour audiobook synced to its text in 20 minutes</title><link>https://agentsinthewildai.com/cases/279-claude-coded-it-a-60-hour-audiobook-synced-to-its-text-in-20/</link><guid>https://agentsinthewildai.com/cases/279-claude-coded-it-a-60-hour-audiobook-synced-to-its-text-in-20/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>For syncing epub and audiobook files, they had Claude implement forced alignment using a CTC transcription model. It aligns a 60 hour book's transcript in 20 minutes, even with dirty source data.</description></item><item><title>Faster than any other Mac app: the markdown viewer Claude Code built</title><link>https://agentsinthewildai.com/cases/278-faster-than-any-other-mac-app-the-markdown-viewer-claude-cod/</link><guid>https://agentsinthewildai.com/cases/278-faster-than-any-other-mac-app-the-markdown-viewer-claude-cod/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>Markdown files should just open, so they asked Claude Code for a viewer. It opens instantly: faster, they say, than any other macOS app, and faster than Bitwarden's Chrome extension shows its fill menu.</description></item><item><title>A routine password reset migration cost about 2 months of Claude Pro</title><link>https://agentsinthewildai.com/cases/277-a-routine-password-reset-migration-cost-about-2-months-of-cl/</link><guid>https://agentsinthewildai.com/cases/277-a-routine-password-reset-migration-cost-about-2-months-of-cl/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>Migrating a password reset flow, Claude Code and Opus produced code on par with a mid level developer. It needed close supervision throughout, and cost roughly 2 months of Pro subscription for about half a day's work.</description></item><item><title>Claude decided on its own to build pixel-by-pixel tests for Quake</title><link>https://agentsinthewildai.com/cases/276-claude-decided-on-its-own-to-build-pixel-by-pixel-tests-for/</link><guid>https://agentsinthewildai.com/cases/276-claude-decided-on-its-own-to-build-pixel-by-pixel-tests-for/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>Working on a Quake project, Claude decided on its own to build a test suite: demo playbacks with snapshots compared pixel by pixel, plus recorded demos whose screenshots it checked for each element.</description></item><item><title>No agency, no screen recording: Claude Code made the launch video</title><link>https://agentsinthewildai.com/cases/275-no-agency-no-screen-recording-claude-code-made-the-launch-vi/</link><guid>https://agentsinthewildai.com/cases/275-no-agency-no-screen-recording-claude-code-made-the-launch-vi/</guid><pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate><description>A side project needed a launch video, but not an agency's bill. Claude Code built one from the codebase plus their feedback. It took a lot of iteration over 2-3 days, all on subscription token usage.</description></item><item><title>ChatGPT checked Kayak and found an earlier way home from Costa Rica</title><link>https://agentsinthewildai.com/cases/274-chatgpt-checked-kayak-and-found-an-earlier-way-home-from-cos/</link><guid>https://agentsinthewildai.com/cases/274-chatgpt-checked-kayak-and-found-an-earlier-way-home-from-cos/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Booking flights for a Costa Rica trip, a picky traveler asked ChatGPT to look them up. It checked Kayak, noticed the return flights got in too late and offered other options; they ended up booking one.</description></item><item><title>Grok Bot now summarizes every call for a home service business</title><link>https://agentsinthewildai.com/cases/273-grok-bot-now-summarizes-every-call-for-a-home-service-busine/</link><guid>https://agentsinthewildai.com/cases/273-grok-bot-now-summarizes-every-call-for-a-home-service-busine/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A home service business connected Grok Bot to its phone system after a few hours of setup. It now listens to every call and posts a summary to the team's Slack channel.</description></item><item><title>Taxes done twice, with Claude and TurboTax, reconciled to the dollar</title><link>https://agentsinthewildai.com/cases/272-taxes-done-twice-with-claude-and-turbotax-reconciled-to-the/</link><guid>https://agentsinthewildai.com/cases/272-taxes-done-twice-with-claude-and-turbotax-reconciled-to-the/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A former PwC tax preparer did their own return with Claude, then again in TurboTax. Claude read the forms and statements and caught a missed credit, and the two matched to the dollar after reconciling.</description></item><item><title>Claude Code helped move a SaaS from AWS to DigitalOcean in 4 hours</title><link>https://agentsinthewildai.com/cases/271-claude-code-helped-move-a-saas-from-aws-to-digitalocean-in-4/</link><guid>https://agentsinthewildai.com/cases/271-claude-code-helped-move-a-saas-from-aws-to-digitalocean-in-4/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A URL shortener side project ran on AWS. Claude Code wrote the GitHub Actions workflows for the move and fixed each failure as it came up. The writer reports zero downtime.</description></item><item><title>Claude Code did the edits; a monorepo's checks fell from 29s to 6.36s</title><link>https://agentsinthewildai.com/cases/270-claude-code-did-the-edits-a-monorepo-s-checks-fell-from-29s/</link><guid>https://agentsinthewildai.com/cases/270-claude-code-did-the-edits-a-monorepo-s-checks-fell-from-29s/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A 7-app monorepo needed an 8 GB memory setting and 25 seconds to typecheck. Claude Code made the changes for the move to oxlint, oxfmt and tsgo, and the team validated each step.</description></item><item><title>Cowork built a one-screen KPI dashboard from 12 tabs in 20 minutes</title><link>https://agentsinthewildai.com/cases/269-cowork-built-a-one-screen-kpi-dashboard-from-12-tabs-in-20-m/</link><guid>https://agentsinthewildai.com/cases/269-cowork-built-a-one-screen-kpi-dashboard-from-12-tabs-in-20-m/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Every Monday a 12-tab marketing tracker got a 30-minute scroll and a vague gut check. Claude Cowork read every tab and built a color-coded dashboard inside the live Google Sheet.</description></item><item><title>Cursor on a data science project: great at Streamlit, weak at pandas</title><link>https://agentsinthewildai.com/cases/268-cursor-on-a-data-science-project-great-at-streamlit-weak-at/</link><guid>https://agentsinthewildai.com/cases/268-cursor-on-a-data-science-project-great-at-streamlit-weak-at/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A data scientist tried Cursor on a real project. Prototyping, code completion and commenting went well, but pandas code kept failing and it deleted a bunch of data files.</description></item><item><title>Claude scrapes Substack rankings weekly; Cowork writes the report</title><link>https://agentsinthewildai.com/cases/267-claude-scrapes-substack-rankings-weekly-cowork-writes-the-re/</link><guid>https://agentsinthewildai.com/cases/267-claude-scrapes-substack-rankings-weekly-cowork-writes-the-re/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Tracking which newsletters break into Substack's top 200 meant clicking and copy-pasting. Claude in Chrome now scrapes the charts every Tuesday, and Cowork turns the data into a spreadsheet report.</description></item><item><title>No way to schedule WhatsApp texts, so Claude Code built one</title><link>https://agentsinthewildai.com/cases/266-no-way-to-schedule-whatsapp-texts-so-claude-code-built-one/</link><guid>https://agentsinthewildai.com/cases/266-no-way-to-schedule-whatsapp-texts-so-claude-code-built-one/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>WhatsApp has no native way to schedule messages, and the writer didn't want third-party apps in their chats. Claude Code built a scheduler using a headless browser, and a recurring message now goes out every morning.</description></item><item><title>Every night, ChatGPT checks orders and drafts the invoice</title><link>https://agentsinthewildai.com/cases/265-every-night-chatgpt-checks-orders-and-drafts-the-invoice/</link><guid>https://agentsinthewildai.com/cases/265-every-night-chatgpt-checks-orders-and-drafts-the-invoice/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Completed orders needed checking against CRM records before billing. Every night ChatGPT reviews each order form against the CRM for mistakes, then creates a draft invoice in the accounting system for review.</description></item><item><title>Claude Code combed credit card statements and saved hours</title><link>https://agentsinthewildai.com/cases/264-claude-code-combed-credit-card-statements-and-saved-hours/</link><guid>https://agentsinthewildai.com/cases/264-claude-code-combed-credit-card-statements-and-saved-hours/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Possible business expenses were buried in a stack of credit card statements. Claude Code went through every statement, flagged the charges that might need reimbursing and put them in an Excel sheet.</description></item><item><title>OpenClaw turned a weekend of hospital forms into a lunch break</title><link>https://agentsinthewildai.com/cases/263-openclaw-turned-a-weekend-of-hospital-forms-into-a-lunch-bre/</link><guid>https://agentsinthewildai.com/cases/263-openclaw-turned-a-weekend-of-hospital-forms-into-a-lunch-bre/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Hospital paperwork needed sorting, some in a language still being learned. An OpenClaw agent scanned 34 pages, flagged missing signatures and drafted an insurance pre-approval for review.</description></item><item><title>Claude Code archived their LinkedIn posts to their blog</title><link>https://agentsinthewildai.com/cases/262-claude-code-archived-their-linkedin-posts-to-their-blog/</link><guid>https://agentsinthewildai.com/cases/262-claude-code-archived-their-linkedin-posts-to-their-blog/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Wanting a backup of their own LinkedIn posts before the platform's bot-blocking caught up, the writer had Claude Code drive a headless browser through their feed and copy the posts to their personal blog.</description></item><item><title>A 3.5-7 hour client audit, done by Claude in 0.5 hours</title><link>https://agentsinthewildai.com/cases/261-a-3-5-7-hour-client-audit-done-by-claude-in-0-5-hours/</link><guid>https://agentsinthewildai.com/cases/261-a-3-5-7-hour-client-audit-done-by-claude-in-0-5-hours/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A technical SEO and UX audit for a new client site normally takes the writer 3.5 to 7 hours. Claude pulled client info from OneDrive, ran the audit, drove a remote PageSpeed tool, and produced a Word report in 0.5 hours.</description></item><item><title>Aider built a $27.81 bridge between Amazon and YNAB</title><link>https://agentsinthewildai.com/cases/260-aider-built-a-27-81-bridge-between-amazon-and-ynab/</link><guid>https://agentsinthewildai.com/cases/260-aider-built-a-27-81-bridge-between-amazon-and-ynab/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Matching Amazon orders to YNAB entries by hand was tedious, so the writer vibe-coded a Chrome extension with Aider that reads Amazon orders, matches them to YNAB, and uses Claude's Haiku to clean up the memo line.</description></item><item><title>Codex built each AWS design in a day, not an estimated three or four</title><link>https://agentsinthewildai.com/cases/259-codex-built-each-aws-design-in-a-day-not-an-estimated-three/</link><guid>https://agentsinthewildai.com/cases/259-codex-built-each-aws-design-in-a-day-not-an-estimated-three/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Building their first AWS project by vibe coding, the writer designed the architecture and let Codex write the code. A flawed first design took a day, against an estimated three or four; the redesign also took a day and scaled.</description></item><item><title>Claude built a proof checker, then found its own flaws</title><link>https://agentsinthewildai.com/cases/258-claude-built-a-proof-checker-then-found-its-own-flaws/</link><guid>https://agentsinthewildai.com/cases/258-claude-built-a-proof-checker-then-found-its-own-flaws/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>After building a minimal proof assistant, the writer asked Claude how they knew the prover was actually sound. Claude checked, found five different ways it could prove false theorems, and fixed them.</description></item><item><title>Never built an extension, Claude made them a Firefox chats organizer</title><link>https://agentsinthewildai.com/cases/257-never-built-an-extension-claude-made-them-a-firefox-chats-or/</link><guid>https://agentsinthewildai.com/cases/257-never-built-an-extension-claude-made-them-a-firefox-chats-or/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>An unwieldy pile of Claude chats sent the writer looking for an organizer extension, but the only one they found was Chrome-only. Claude built one from scratch for Firefox instead.</description></item><item><title>Three AI agents raced through a 50-page tax return</title><link>https://agentsinthewildai.com/cases/256-three-ai-agents-raced-through-a-50-page-tax-return/</link><guid>https://agentsinthewildai.com/cases/256-three-ai-agents-raced-through-a-50-page-tax-return/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A roughly 50-page tax return, run through three coding agents side by side. Codex was accurate first time in 12 minutes, Antigravity missed a value until prompted, and Claude Code stumbled overnight.</description></item><item><title>Codex and Claude finally built the RSS reader they shelved</title><link>https://agentsinthewildai.com/cases/255-codex-and-claude-finally-built-the-rss-reader-they-shelved/</link><guid>https://agentsinthewildai.com/cases/255-codex-and-claude-finally-built-the-rss-reader-they-shelved/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Four years after shelving a homemade RSS reader and selling off its $500 domain, the writer built it this year with Codex and Claude. They use it daily, along with 30 other strangers who found it public.</description></item><item><title>Claude rebuilt a 1996 toaster screensaver, with pushback</title><link>https://agentsinthewildai.com/cases/254-claude-rebuilt-a-1996-toaster-screensaver-with-pushback/</link><guid>https://agentsinthewildai.com/cases/254-claude-rebuilt-a-1996-toaster-screensaver-with-pushback/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Chasing a childhood itch for the exact 1996 After Dark Flying Toasters screensaver, the writer turned to Claude. It did a lot of good reverse-engineering work, but kept inventing details that needed pushback.</description></item><item><title>Hermes Agent kept watch and booked the swim lesson</title><link>https://agentsinthewildai.com/cases/253-hermes-agent-kept-watch-and-booked-the-swim-lesson/</link><guid>https://agentsinthewildai.com/cases/253-hermes-agent-kept-watch-and-booked-the-swim-lesson/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Booking swimming lessons for their son was on the to-do list. Hermes Agent monitored availability and automatically booked a suitable slot the moment one opened up.</description></item><item><title>Gemini curates a twice-weekly news digest email</title><link>https://agentsinthewildai.com/cases/252-gemini-curates-a-twice-weekly-news-digest-email/</link><guid>https://agentsinthewildai.com/cases/252-gemini-curates-a-twice-weekly-news-digest-email/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>To keep up with industry and tech news, the writer built a free-tier Google Cloud Run job pulling Google News RSS feeds. Gemini picks which articles match their criteria and triggers a summary email.</description></item><item><title>Claude Cowork sized a 10,000-node cluster in 30 seconds</title><link>https://agentsinthewildai.com/cases/251-claude-cowork-sized-a-10-000-node-cluster-in-30-seconds/</link><guid>https://agentsinthewildai.com/cases/251-claude-cowork-sized-a-10-000-node-cluster-in-30-seconds/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>During a live call, the writer dumped nomad-job-inspect data for a 10,000-node cluster into Claude Cowork and asked for memory/CPU requirements by job, datacenter and SKU. It built the table in under 30 seconds.</description></item><item><title>ChatGPT ported a Chrome plugin to Safari in 20 minutes</title><link>https://agentsinthewildai.com/cases/250-chatgpt-ported-a-chrome-plugin-to-safari-in-20-minutes/</link><guid>https://agentsinthewildai.com/cases/250-chatgpt-ported-a-chrome-plugin-to-safari-in-20-minutes/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Safari's lack of a favorite Chrome plugin was the writer's last complaint about it. ChatGPT pulled the source from their installed Chrome extension, ported it, and installed it in Safari in 20 minutes.</description></item><item><title>OpenClaw scans used bookstores against a Goodreads list</title><link>https://agentsinthewildai.com/cases/249-openclaw-scans-used-bookstores-against-a-goodreads-list/</link><guid>https://agentsinthewildai.com/cases/249-openclaw-scans-used-bookstores-against-a-goodreads-list/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>Wanting their son to grow up around books, they hooked OpenClaw to their Goodreads list to scan local used bookstores for titles to buy. They say they've never read as much or enjoyed it as much as now.</description></item><item><title>Claude wrote a test harness for a bug no one wanted to chase</title><link>https://agentsinthewildai.com/cases/248-claude-wrote-a-test-harness-for-a-bug-no-one-wanted-to-chase/</link><guid>https://agentsinthewildai.com/cases/248-claude-wrote-a-test-harness-for-a-bug-no-one-wanted-to-chase/</guid><pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate><description>A rabbitmq connector occasionally got stuck after rare network drops, too minor to spend days fixing. Claude built a test harness that reproduced it, and the actual fix wasn't too hard.</description></item><item><title>Hermes Agent wrote, sang and made a whole 8D music video</title><link>https://agentsinthewildai.com/cases/247-hermes-agent-wrote-sang-and-made-a-whole-8d-music-video/</link><guid>https://agentsinthewildai.com/cases/247-hermes-agent-wrote-sang-and-made-a-whole-8d-music-video/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>An 8D music video with millions of views sparked a test: could an agent make one? Hermes Agent wrote the lyrics and music, made visuals from the song and added the 8D effect and audio-responsive video.</description></item><item><title>Claude Cowork turned a meal-prep video into a stocked grocery cart</title><link>https://agentsinthewildai.com/cases/246-claude-cowork-turned-a-meal-prep-video-into-a-stocked-grocer/</link><guid>https://agentsinthewildai.com/cases/246-claude-cowork-turned-a-meal-prep-video-into-a-stocked-grocer/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>Watching a meal-prep video and building a shopping list by hand ate up time each week. Claude Cowork extracted the recipes, matched ingredients to past-invoice brands, filled the cart, and saved the recipes to Notion.</description></item><item><title>Devin's first PR took 10 minutes. Its 150 credits lasted under a week</title><link>https://agentsinthewildai.com/cases/245-devin-s-first-pr-took-10-minutes-its-150-credits-lasted-unde/</link><guid>https://agentsinthewildai.com/cases/245-devin-s-first-pr-took-10-minutes-its-150-credits-lasted-unde/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>A client's CRM needed a new deals feature. Devin opened a first pull request in under 10 minutes, but its PRs carried unneeded changes, it gave a false answer about a library, and it hit its usage limit.</description></item><item><title>OpenClaw read 15+ pull requests to recommend an upgrade</title><link>https://agentsinthewildai.com/cases/244-openclaw-read-15-pull-requests-to-recommend-an-upgrade/</link><guid>https://agentsinthewildai.com/cases/244-openclaw-read-15-pull-requests-to-recommend-an-upgrade/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>When a new OpenClaw beta added Telegram message streaming, they asked their agent what else had changed. It read 15+ merged pull requests, compared them to the current version, and gave an upgrade recommendation.</description></item><item><title>Claude applied for 50 jobs in less than 30 minutes</title><link>https://agentsinthewildai.com/cases/243-claude-applied-for-50-jobs-in-less-than-30-minutes/</link><guid>https://agentsinthewildai.com/cases/243-claude-applied-for-50-jobs-in-less-than-30-minutes/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>Applying to jobs one by one eats hours. Using the new Claude Computer Use, 50 job applications went out in less than 30 minutes.</description></item><item><title>40 strangers, one conference: Claude Code built the briefing</title><link>https://agentsinthewildai.com/cases/242-40-strangers-one-conference-claude-code-built-the-briefing/</link><guid>https://agentsinthewildai.com/cases/242-40-strangers-one-conference-claude-code-built-the-briefing/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>Before a conference with 40 people they had never met, LinkedIn research normally takes an evening. Claude Code researched the attendees in parallel via LinkedIn and Chrome MCP, producing a full briefing in 5 minutes.</description></item><item><title>A homemade agent turns vet charges into ready-to-send insurance claims</title><link>https://agentsinthewildai.com/cases/241-a-homemade-agent-turns-vet-charges-into-ready-to-send-insura/</link><guid>https://agentsinthewildai.com/cases/241-a-homemade-agent-turns-vet-charges-into-ready-to-send-insura/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>Vet bills mean a paper chase of card charges, emailed invoices and insurer forms. A homemade agent matches each charge to its Gmail invoice, fills the claim form with Gemini's help and stages a draft. A person hits send.</description></item><item><title>A Raspberry Pi car assistant answers by voice, no internet needed</title><link>https://agentsinthewildai.com/cases/240-a-raspberry-pi-car-assistant-answers-by-voice-no-internet-ne/</link><guid>https://agentsinthewildai.com/cases/240-a-raspberry-pi-car-assistant-answers-by-voice-no-internet-ne/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>A Raspberry Pi rides in the car and runs a local AI model that answers spoken questions through the car's own speakers, even with the internet switched off. It also posts engine faults into the household chat room.</description></item><item><title>Played thinking aloud, then Claude checked each thought with Stockfish</title><link>https://agentsinthewildai.com/cases/239-played-thinking-aloud-then-claude-checked-each-thought-with/</link><guid>https://agentsinthewildai.com/cases/239-played-thinking-aloud-then-claude-checked-each-thought-with/</guid><pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate><description>A recorded rapid chess game came with a voice memo of what the player was thinking move by move. Claude drove Stockfish to question each thought and built a narrated video explaining the mistakes.</description></item></channel></rss>