A little less work admin.

Scanned emails for interests and portfolio holdings to populate the Muse feed.

Ran Claude Code on Muse’s Linux VM (2 vCPU, ~8GB RAM, 100GB disk) via a tiny outbound tunnel.

Built a basic CRM, tiered leads, and spun up a sub-agent to handle outreach.

Prospected 50 ICP accounts, built an enriched CSV, and drafted initial outreach.

Took a Google Analytics report and rewrote figures, charts, and details after Claude and Gemini refused.

Audited an entire YouTube channel end to end.

Migrated multi-agent Claude workflows to Muse for daily morning runs, cleaned junk email, and cancelled found subscriptions.

"one of the tasks i get muse to do it : find xx firms or agency in xx region based on my {a b c d} requirements, and get their background and their previous work. it found several candidates and emailed them all on my behalf. wild."

Automates a professional Instagram with post context, captions, audio, editing, hashtags, and scheduled reels.

Researches and procures prototype components from many different suppliers.

Travel photographer The user's agent "Rube" built their content calendar from Instagram/Threads/LinkedIn, reconciled invoices, booked two Top of the Rock sunset tickets in NYC via the agent's browser, and organized 308 saved Instagram locations into an interactive destination map. (Sponsored…

Runs Shopify stores, handles email, coordinates with employees, and manages Asana.

A Muse user built humanbrain on tribe v2 this week : Muse Spark 1.3 wrote basically all of it. It asked two clarifying questions, then shipped in 3 days what they'd budgeted a month for. It even caught a bug in their own preprocessing before they did.

A Muse user was interviewing for a company and had Muse analyze all their active Meta ads, broken out by creative.

Scanned edited-video folders for posts that never went live and scheduled them all in about two minutes.

A Muse user was down childcare for their youngest (8 months) and on dad duty all day. With Muse's help they launched two websites from the ground up in a week, updated them significantly, and sent out cold email pitches for one of them.

A Muse user had their Muse build them an app that finds relevant jobs and provides Q&A on gaps between the posting and their resume.

A Muse user fed 30 years of unsorted financials to their model and told Muse which cells needed filling : it did it perfectly every time, across companies that break out their financials differently.

A Muse user connected their credit cards via Plaid to build a business-travel expense dashboard. They mentioned a previous business trip as an example, and Muse went through their transaction history and automatically pulled all the expenses.

A Muse user turned their Muse into a foreman named Bart: they scouts stories for their X account, handles real admin work, spins up subagents when a job needs extra hands, and challenges their own reasoning. They's an ironworker learning AI by giving Bart real jobs instead of reading a feature list.

A Muse user built BillRx, an AI agent that audits confusing medical bills : from a one-word prompt to a submitted hackathon entry in about 3 hours: working prototype, repo, demo video. The model never touches the money math; that's all deterministic code.

A Muse user is using Muse to build a sample website for a local dog groomer : praising how it confirms understanding and fixes small details quickly.

A Muse user has Muse “stuff working 24/7 in the background for my business and my family” on the free tier without touching usage limits.

A Muse user asked Muse to go to YouTube, find a Zuckerberg video about Muse, cut the specific part, add burned-in subtitles, render it, and deliver an X-ready video : “it did the whole thing.”