Every entry links to its source, rates the evidence and lists what went wrong.
75 use cases
Every week, pulls new videos from a list of YouTube channels, rewrites their transcripts as magazine articles, and emails them to you as a newsletter plus an EPUB file.
Isolated cron jobs scan Reddit, check inboxes, and pull Google Search Console data, then send summaries to Telegram.
The agent researches topics and writes blog posts in Markdown. The human reviews and publishes, saving 3 to 4 hours a week across two sites.
A native, clearly labelled card in the directory results.
Become the featured sponsor →An agent running 24/7 on a Mac Mini finds prospects, verifies emails, sends outreach, makes PDF briefings, and tracks the pipeline, keeping a lessons log so it improves over time.
Claude Code sends a pile of videos, PDFs and articles through NotebookLM and turns the output into linked Obsidian notes covering scenarios, actors and probabilities.
The agent sets up an ElevenLabs voice agent connected to its own gateway and adds a talk widget to the website.
Within 2 hours, Claude Cowork answered 42 Telegram DMs asking for intros, revised 3 pitch decks, and reviewed 9 drafts with inline comments.
A morning brief from Google Drive context arrives on Telegram between 7 and 8am. The same author says OpenClaw never got it right in a week of trying.
A HubSpot webhook alerts a 'Researcher' agent, which scrapes the prospect's site and LinkedIn for recent news. The trigger itself is a plain webhook, not an LLM.
An ops workflow handles inbox triage, quick research, and a weekly summary, with a 'supervisor' agent that enforces a checklist (sources, confidence, next actions) and stops hard when tools fail.
Claude Code checked every video against a year of real search queries. It rewrote 23 titles and found more than 100 videos linking to dead links.
Pulls what people actually search on the channel, drafts the title and description from that, generates thumbnails, and schedules the best one.
A single SKILL.md holds schedules, dietary profiles and the custody calendar. It splits the grocery list across 4 stores by price and fills carts via Chrome MCP.
Merges work, family and partner calendars, OCRs school calendar PDFs, scans email for invites, and sends one morning briefing.
An IMAP trigger feeds an OpenAI classifier that routes 80–120 emails a day to Slack, Pipedrive and Zendesk. It hit 94% accuracy on 500 emails and saves 11 hours a week.
Part of a 15-cron home setup. It triages Gmail every hour (labels actionable items, archives noise) and emails a daily briefing with weather, calendar, system stats and tasks.
An Elvis tribute singer says Grok Bot emails venues, negotiates prices, closes deals and reports back, and booked 30 shows in a month.
Same research-and-save task given to three agents. Comet used the existing in-browser Notion session and created the note in one go.
A non-engineer built a 'boring' vertical B2B SaaS with an orchestrator chat plus worker chats. Two paying clients and 6 leads, about $800 in credits spent.
Shipped in about a month on agent-friendly tooling (Next.js, Convex, Clerk, Stripe, Vercel). It was free for 3 months, and growth came entirely from long-tail SEO landing pages.
A shared Gmail label feeds an extractor, totals are validated to €0.01, low-confidence fields go to a human, and results push to Sage. 87% need no human touch.
n8n handles document management and client chasers, letting a bookkeeping firm quadruple clients without hiring.
Each day, finds a historical event from Wikipedia's 'On This Day', generates an image of the moment 10 seconds before it, and posts it to Discord as a guessing game.
Scrapes 500+ leads, verifies the emails, writes a personalized icebreaker from each lead's website, and uploads everything to an Instantly campaign.
Send a task over Telegram, the agent writes the code and opens a PR, and a nightly job has the agent review its own logs and improve.
Scheduled jobs send a daily brief covering mail, calendar and tasks, plus a weekly meal plan.
The agent orders groceries from the Rohlik store through its MCP endpoint, using the weekly meal plan.
A single always-on agent runs blog publishing, email drips via Resend, social engagement, analytics, a quiz funnel, ad monitoring, and daily heartbeat digests.
The agent publishes static sites to surge.sh as needed. The main use is a personal news and trends dashboard.
Before a client call, the agent built a meeting-prep website and a demo dashboard of the data it can pull, tested on a real use case.
An agent drafted 21 emails to high-profile investors on its own without asking for final permission. The drafts were good, but it showed why approval gates matter.
The creator trained a Claude system on his 20 best scripts and his own research process (browsing for outliers via Chrome MCP), and it produced a video with 152K views.
A creator A/B tests 3 titles and then 3 thumbnails, and a ChatGPT-suggested, Nano Banana-rendered thumbnail outperformed both of his originals.
A cron job scouts trending topics, checks them against past pitches using embeddings to avoid repeats, and pitches new ideas with sources to a Telegram topic.
Researches the guest and topic, writes a cold-open hook, intro, and 5–7 questions from easy to provocative, then produces show notes and promo posts.
Describe edits in plain language (trim, merge, music, subtitles, color, 9:16 crop) and the agent applies them with no timeline editor.
Ten free skills built on the YouTube Data API cover a research-to-publish loop: niche research, trend scan, competitor analysis, own-channel audit, title/tag optimizer, thumbnail analyzer, and comment miner.
Claude plans the week's meal prep every Sunday and launches Chrome to fill the online cart for curbside pickup. A separate daily job checks store ads for deals.
Claude set up a SimpleFIN bridge and reconciles spending against the budget every day using scheduled tasks on the user's computer.
The agent tested coupon codes at checkout until it found ones that worked, saving $79.
One prompt produced a hosted comparison site for the Yamaha YDP-141 and Casio AP270, with pros, cons and review summaries, plus a Turkish language toggle.
Manus planned a 2-week itinerary covering post-Brexit pet paperwork, Eurotunnel vs ferry, and dog-friendly hotels, activities and restaurants.
The same flight request (LGA/JFK to SJC/SFO, Delta One preferred) went to 4 agents. Grok Bot gave ranked options and booked the flight itself.
In under an hour after install, Grok Bot was working on an Enterprise refund, booked a Christmas flight, and found recliner deals.
A personal agent on a local model found an in-person piano teacher, and the poster is now taking lessons.
Check-ins at set times for each habit, with streaks tracked in a local file and a tone that adapts to progress.
Log meals and symptoms by message, get 3 daily reminders, and receive a weekly analysis of which foods line up with symptoms.
Every night it reads the last 24 hours of newsletters, sends a digest of the important bits with links, and asks for feedback to improve.
Finds and tracks contacts from email and calendar (first seen, last contact, count, notes) and answers plain-language questions about them.
Paste a transcript and get a 5-bullet summary plus every action item with owner and deadline. It creates Jira tickets and posts to #meeting-notes.
Calls each guest on a list to confirm attendance, collects notes, and compiles a summary.
A user spent 1.5 hours supervising Manus on 15 minutes of booking and calendar work. Neither Manus nor ChatGPT Agent could create calendar events from Gmail travel emails, but Claude chat did it first try in under a minute.
Picks a contact from the database, writes a briefing, drafts a custom message, attaches portfolio assets, builds a presentation, logs it in the CRM, sends, follows up, and adds booked meetings to Calendar. It runs daily.
A context bug made the agent drop its polite template and send blunt, specific critiques of founders' sites to 40 founders. The author reports a 62% reply rate, 8 calls, and 2 paid $300 audits.
Sales Navigator list enriched in Clay, Claude writes a personalized first email, 2 follow-ups and 3 subject lines, then emails are verified and sent via Smartlead.
A builder posted plain n8n workflow-run videos on a 40-subscriber channel, and a viewer hired him for €750 paid upfront.
Every day at 5pm it sends the top posts from chosen subreddits, asks whether you liked them, and saves your preferences as rules.
Every Sunday it posts the week's tech earnings calendar. You pick companies, it schedules a job per report, and posts beat/miss, EPS, guidance and AI highlights.
Before coding anything, it scans GitHub, HN, npm, PyPI and Product Hunt for competitors and stops if the space is crowded.
Drop URLs, tweets, YouTube links or PDFs in chat to ingest them, then ask questions and get ranked excerpts with sources. Other workflows query it too.
An amateur used Claude Code to search public data and flagged what may be an undiscovered planet. It reached 301 points on HN.
Paper trades on prediction markets with backtesting, strategy analysis and daily P&L reports. No real money.
An MCP server gets the top 10 niche videos with engagement metrics and thumbnails, then hands off to one subagent that writes a research report to disk.
A former trainer prototyped in Lovable and finished in Cursor. AI workout plans for coaches reached $2,320 MRR, but Claude API costs are $850 a month.
A solo founder used Cursor with Next.js, Supabase and Vercel. Reddit posts brought about 200K views, 4K visits, 385 signups and roughly $80 MRR.
A solo dev shipped a full-stack app with 3 LLM integrations. Scaffolding was about 4x faster, but the AI invented model IDs and parameters, so the rule became checking every external call against the docs.
An always-on agent with SSH, kubectl and 1Password checks health, analyzes logs, fixes pods and configs, deploys through infrastructure-as-code, and runs a nightly brainstorm.
Gave Perplexity Comet step-by-step manual test cases for a live site. It ran each one and reported results.
Code handles predictable invoice lines, Claude handles tricky negative items, structured parsers output to Sheets, and project names are fuzzy-matched to an internal API.
A Tally booking form triggers a webhook, a switch picks the invoice template, and it fills customer and property details, sends the invoice plus a thank-you email, and saves a copy.
Reviewed Every's P&L and weekly performance tracker and produced a PowerPoint on the state of the business.
The agent matched purchase orders to delivery documents and invoices in a Drive folder, tracking progress and variances in a .md file.
WhatsApp, Instagram, email and Google Reviews go into one inbox. It classifies intent, answers FAQs, books appointments, flags complaints and thanks reviewers.
A solo founder runs four agents (strategy lead 'Milo', business, marketing, dev), each with its own SOUL.md, all reached from one Telegram chat.
Draft in Obsidian while the agent co-edits. It makes a 16:9 banner with Gemini, publishes through the Ghost API, and redeploys the static site.
Nothing matches yet. or submit one.