Peter Yang Deploys 11 Grok Bots to Automate Creator Workflow

Peter Yang

Delegating daily operations to specialized, communicating AI agents eliminates creator burnout by running email, research, and personal logistics autonomously.

AI-assisted summary of 11 Grok Bots I Use Every Day (After 100+ Hours) — original source. Check the original for full context.

Source published: 2026-09-14

Key ideas

Rather than relying on an all-in-one assistant, modular personal AI setups operate best through specialized single-purpose bots collaborating in shared channels like a real-world team. (0:00)
By shifting to automated recurring schedules, bots proactively deliver weekly content pitches, fitness logs, flight monitoring, and school briefs without requiring manual prompts. (19:26)
A two-tier leadership model—pairing a Chief of Staff (Chief) for daily email triage and routing with a long-term strategic Advisor—prevents low-level operational noise from derailing high-level goals. (3:49)
Grok Bot agents utilize cloud-computer browser automation rather than expensive official APIs to execute tasks like monitoring X, tracking Google Flights, and listing items on Facebook Marketplace. (9:45)

Detailed notes

## Background Published on YouTube by Peter Yang on September 14, 2026, this video details his refined lineup of 11 daily Grok Bots developed after logging over 100 hours with the platform. According to the video description, the guide demonstrates how these specialized bots handle email, long-term planning, content creation, finances, fitness, and family logistics while collaborating like a real-world team.

## TL;DR Peter Yang breaks down his modular ecosystem of 11 specialized Grok Bots divided into meta-management, business production, and personal life operations. System architects like Dr. Light and Chief handle bot generation, inbox triage, and task delegation, while analytical bots like Advisor and YouTube Producer audit blind spots and draft content packages. On the personal side, dedicated agents manage flight alerts, Facebook Marketplace negotiations, school newsletters, and fitness tracking. Rather than relying on a single monolithic assistant, the setup succeeds through hyper-specialized single-purpose agents that run proactive background schedules and communicate collaboratively in shared channels.

## Atmosphere & style The presentation is pragmatic, energetic, and candidly conversational, feeling less like a dry tech tutorial and more like a behind-the-scenes walkthrough from an experienced founder refining his personal operating system. Yang balances tactical software demonstrations with playful humor, especially when showcasing the bots bantering and debating their relative worth to his life. The video carries an encouraging, builder-first vibe that demystifies multi-agent AI orchestration into accessible, daily habits.

## Bulletin - [00:00](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=0s) Rather than relying on an all-in-one assistant, modular personal AI setups operate best through specialized single-purpose bots collaborating in shared channels like a real-world team. - [19:26](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1166s) By shifting to automated recurring schedules, bots proactively deliver weekly content pitches, fitness logs, flight monitoring, and school briefs without requiring manual prompts. - [03:49](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=229s) A two-tier leadership model—pairing a Chief of Staff (Chief) for daily email triage and routing with a long-term strategic Advisor—prevents low-level operational noise from derailing high-level goals. - [09:45](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=585s) Grok Bot agents utilize cloud-computer browser automation rather than expensive official APIs to execute tasks like monitoring X, tracking Google Flights, and listing items on Facebook Marketplace.

## Key takeaways - [19:26](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1166s) Organizing AI agents into specialized single-purpose roles preserves clean context windows and prevents the quality degradation typical of all-in-one chatbots. - [00:00](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=0s) Creating shared channels where bots introduce themselves and debate trade-offs turns a collection of isolated prompts into a cohesive, interactive workspace. - [02:24](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=144s) The meta-bot Dr. Light acts as an internal systems architect that rapidly drafts prompts, configures tool permissions, and audits underperforming agents. - [03:49](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=229s) Chief functions as an operational router, scrubbing personal data from inboxes, drafting responses, and delegating specific tasks to specialized downstream bots. - [04:59](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=299s) The Brutally Honest Advisor prompt cuts through comfort and blind spots by forcing the AI to highlight wasted time and strategic errors across connected multi-platform data. - [07:08](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=428s) The YouTube Producer automates content pre-production by detecting channel performance outliers, pitching topics, drafting hooks, and pushing packaging ideas directly into Linear tickets. - [08:44](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=524s) Behind the Growth tracks Vercel metrics, conversion funnels, and copy A/B tests to provide clear commercial reporting without requiring direct dashboard logins. - [09:45](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=585s) X Scout relies on a cloud browser to bypass costly APIs, pulling trending AI posts and surfacing forgotten daily bookmarks to eliminate mindless doom-scrolling. - [11:05](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=665s) Frugal Dad monitors complex flight routes and actively negotiates Facebook Marketplace sales, while maintaining a firm security boundary by barring access to bank accounts and Venmo. - [13:21](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=801s) Weekend Planner uses an evolving markdown taste profile built from personal IMDb ratings to recommend narrative-driven sci-fi media while avoiding duplicate suggestions. - [15:18](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=918s) Health Coach combines smart scale measurements and workout logs from custom MCP apps to monitor progressive overload and body composition targets. - [16:28](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=988s) Marie Kondo performs digital hygiene audits on Google Drive and recurring subscriptions, requiring explicit confirmation before archiving files or canceling tools. - [17:15](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1035s) Loving Husband digests dense, text-heavy school newsletters into actionable summaries and tracks household errands. - [19:26](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1166s) The transition from reactive prompt-and-response assistants to autonomous recurring task scheduling is the primary driver of compounding AI productivity. - [20:10](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1210s) Grok Bot distinguishes itself in a crowded assistant market through direct multi-agent intercommunication rather than monolithic chat interfaces.

## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=0s) The 11 Grok Bots and Team Communication

After spending over 100 hours and a month experimenting with Grok Bot, the current lineup of 11 daily bots looks completely different from the initial launch setup. Rather than operating as isolated single-purpose chatbots, the entire system functions like an interactive team where agents actively communicate with one another. To illustrate this during video preparation, the YouTube Producer bot set up two dedicated channels with four to five bots in each room, prompting each bot to introduce itself with its name, core responsibilities, and a concrete example of its work.

```mermaid flowchart TD accTitle: Grok Bot Orchestration Ecosystem accDescr: Diagram showing how Dr. Light and Chief orchestrate and coordinate the specialized bots.

subgraph Meta ["System Management"] DL["Dr. Light<br/>(Bot Builder & Auditor)"] Chief["Chief<br/>(Chief of Staff & Orchestrator)"] end

subgraph Work ["Content & Business"] YP["YouTube Producer"] BG["Behind the Growth"] XS["X Scout"] Adv["Brutally Honest Advisor"] end

subgraph Life ["Personal & Home Operations"] FD["Frugal Dad"] LH["Loving Husband"] HC["Health Coach"] WP["Weekend Planner"] MK["Marie Kondo"] end

To test their interaction, the bots were instructed to debate among themselves over who provided the most value to Peter. The resulting banter showcased their distinct personas: - **YouTube Producer** bowed out of the contest immediately, noting it was merely the facilitator that organized the room. - **Frugal Dad** argued for financial supremacy, declaring, "Respect to Loving Husband, but I'm the only bot that helps Peter save money." - **Loving Husband** pushed back with crucial daily logistics, reminding the room that it prevented Peter from missing his kids' school picture day at 7:50 a.m. - **Health Coach**, **Weekend Planner**, and **Marie Kondo** (who manages Google Drive clutter and email hygiene) all chimed in to defend their utility.

While they never reached a consensus on who mattered most, the collaborative back-and-forth illustrates why Grok Bot stands out: it operates less like a sterile AI utility and more like a dedicated corporate Slack workspace populated by AI staff members.

---

[02:24](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=144s) Dr. Light: Building and Maintaining the Bot Team

The foundation of the entire setup is **Dr. Light**, a meta-bot adapted from Grok Bot team member Lauren Tan's publicly available template on the marketplace. Originally shared as "Dr. Eggbot," it was promptly renamed after the heroic creator from the *Mega Man* video games because Dr. Eggman was the villain. Dr. Light functions as an internal systems architect and auditor, designing high-quality agent prompts, configuring necessary tool access and skills, and running scheduled maintenance checks to refine underperforming bots.

To demonstrate how rapidly Dr. Light works, Peter prompts it on the fly to generate a comedy-focused bot. In seconds, Dr. Light creates a dedicated agent named **Punchline**. When asked for its first joke, Punchline delivers a groan-worthy pun: *"I told my suitcase there would be no vacations this year. Now we're not on speaking terms."* While corny enough that it needs prompt tuning to be more surprising or edgy, the exercise shows how effortlessly Dr. Light generates and deploys new specialized agents.

---

[03:49](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=229s) Chief: Email Triage and Delegation

**Chief** serves as the primary Chief of Staff and central router for Peter's entire operating ecosystem. Rather than requiring direct management of every discrete task, Chief handles delegation across other bots: - When Peter needed to sell a pair of old headphones, Chief was instructed to pass the specs to **Frugal Dad** so it could draft and publish the listing on Facebook Marketplace. - Chief coordinates family workflow by scheduling **Loving Husband** to scan school newsletters every weekday at 4:00 p.m.

Beyond task routing, Chief handles traditional executive assistance by parsing incoming emails and calendar schedules. Using its remote computer environment and built-in browser automation, Chief scans Peter's inbox, strips out sensitive personal identifiable information, and presents a summarized action checklist. In one demonstration, Chief identifies a reader offering medical advice, a pending workshop follow-up, and an upcoming party invitation. Peter can simply dictate rapid numbered directives—such as telling it to draft a reply for item one, approve item two, and RSVP to item three—and Chief executes the actions directly in the background.

---

[04:59](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=299s) Advisor: Strategic Planning and Blind-Spot Auditing

While Chief manages low-level daily triage, **Advisor** focuses purely on high-level, long-term strategic direction. The bot has read access to a living Google Doc containing Peter's multi-month goals and receives automated data inputs from connected accounts across Mercury Bank, YouTube analytics, Granola meeting notes, Substack, and the Behind the Craft website.

Every Friday at 8:00 a.m., Advisor delivers an automated strategic briefing that outlines core weekly priorities (such as completing the Personal AI System course), highlights video concepts that are overperforming on competing channels, and flags upcoming meetings requiring prep.

To extract the most value from this setup, Peter uses a dedicated persona instruction known as the **Brutally Honest Advisor prompt**:

> *"Act as my brutally honest advisor.* > > *Speak to me like a founder, creator, or leader with massive potential and blind spots that need to be exposed. I don’t want comfort or fluff. I want the truth, even when it stings.* > > *Analyze my situation objectively. Tell me what I’m doing wrong, underestimating, avoiding, or making excuses about. Show me where I’m wasting time, moving too slowly, or playing small.* > > *Then tell me exactly what I need to change, prioritize, or build to reach the next level. Be direct, specific, and strategic.* > > *If I’m lost or making a mistake, say so and explain why. Hold nothing back, but end on an encouraging note."*

Peter notes that running this prompt against extensive context files consistently uncovers uncomfortable truths and profound strategic corrections that creators typically avoid facing on their own.

---

[07:08](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=428s) YouTube Producer: Video Outliers and Script Packaging

Peter's professional focus centers on creating YouTube videos, publishing newsletters, and testing software. The **YouTube Producer** bot handles the research and pre-production pipeline:

- **Outlier Detection:** The bot automatically scans performance data across Peter's channel as well as peer creator accounts to find videos outperforming rolling 7-day averages. - **Automated Ideation Schedule:** Every Monday and Wednesday at 7:00 a.m., it delivers a curated list of high-potential content topics (which directly led to prioritizing the Grok Bot tutorial series). - **Collaborative Scripting:** The bot builds structural hooks, pulls research examples, and outlines bulleted talking points. - **Project Tracking Integration:** After Peter refines the tone so it sounds like his natural conversational speaking style rather than a stiff script, YouTube Producer bundles the title variations, packaging ideas, thumbnail concepts, and script outlines directly into a ticket on Linear.

---

[08:44](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=524s) Behind the Growth: Funnel Tracking and Conversion Analytics

**Behind the Growth** monitors web analytics and commercial conversions for the behindthecraft.com ecosystem. Pulling data directly from Vercel hosting metrics alongside custom internal conversion tracking, the bot tracks key metrics: - Overall visitor volume and geographic/channel traffic sources. - Reader-to-paid-subscriber conversion rates flowing through the sales funnel. - Performance statistics on active A/B tests testing headline copywriting variations.

The bot automatically generates summary reports (with sensitive revenue figures redacted for demonstrations) and creates comparative performance charts, giving Peter a clear operational picture of audience behavior and paid conversion performance without requiring manual analytics dashboard logins.

[09:45](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=585s) X Scout and Automated Social Monitoring

Moving into social media and content curation, the creator introduces X Scout, a bot that crawls X (formerly Twitter) to pull trending AI tweets and the creator's latest website posts. While using the official X API is an option, it incurs costs, so this bot takes a different approach: it logs into X on a cloud computer and browses like a human to harvest the information. On any given day, it captures discussions from industry leaders like Dario Amodei and Sam Altman regarding frontier model development and human alignment.

Beyond public trends, X Scout addresses a common personal habit: saving bookmarks that never get read. The bot surfaces these neglected bookmarks every morning, running automatically so the creator can stay informed without falling into endless doom-scrolling loops.

- Pulls trending AI tweets from leaders like Dario Amodei and Sam Altman discussing frontier model alignment. - Scrapes the author's latest website posts and surfaces unread bookmarks every morning. - Operates via cloud-computer browser automation to bypass expensive API costs.

---

[11:05](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=665s) Frugal Dad: Finance, Flight Tracking, and Marketplace Automation

Shifting to personal and financial tasks, the Frugal Dad bot manages miscellaneous money-related workflows. One primary use case is monitoring December flight fares to Japan across different legs, such as an open-jaw trip from SFO to Tokyo and Fukuoka back to SFO, versus a standard Tokyo roundtrip. Using Google Flights on its cloud browser, the bot advises whether to buy immediately or wait a few weeks based on falling price trends.

Frugal Dad also automates Facebook Marketplace by listing items—such as a pair of headphones—and handling buyer negotiations according to the creator's pricing rules. The headphones were successfully sold just an hour before recording. While the bot requires access to accounts like Facebook and Google Flights, the creator deliberately avoids connecting banking credentials or Venmo for security.

- Tracks Japan flight fares for December on Google Flights (comparing open-jaw and roundtrip options) and advises on optimal purchase timing. - Lists items on Facebook Marketplace and actively negotiates with buyers on pricing parameters. - Maintains a boundary by avoiding direct access to sensitive financial accounts like bank logins or Venmo.

---

[13:21](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=801s) Weekend Planner and the IMDb Taste Profile

To combat the local Bay Area workaholic culture and find family activities, the Weekend Planner bot sends a concise Saturday and Sunday family brief every Wednesday via email. The brief highlights local events, such as Google's endless summer festival party in Mountain View. Additionally, the bot suggests movies and TV shows tailored to the creator's precise entertainment preferences by reading a custom `taste profile.md` file generated from IMDb ratings.

This taste profile documents the creator's preference for strong characters and narrative momentum in science fiction, alongside an exhaustive list of 10-out-of-10 and lower-rated movies. Whenever the creator watches something new—like recently watching *Toy Story 5* and rating it 9 out of 10—they instruct the bot to update the taste profile, ensuring increasingly refined recommendations and preventing duplicate suggestions.

- Sends a Wednesday email brief containing weekend family activities and tailored movie/TV show recommendations. - Relies on a custom `taste profile.md` file compiled from IMDb ratings, prioritizing strong characters, narrative momentum, and science fiction. - Continuously updated in real time (e.g., adding *Toy Story 5* at 9/10) to improve future suggestions and filter out previously watched content.

---

[15:18](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=918s) Health Coach and Marie Kondo Digital Cleanup

The Health Coach bot delivers a comprehensive report every Saturday at 8 AM, synthesizing data from a Withings smart scale (weight and body fat percentage) and workout logs from a custom fitness app built via vibe coding and Model Context Protocol (MCP). For example, the report noted a weekly volume of 27,000 pounds lifted across five lifting sessions, alongside squat and bench press metrics, while keeping tabs on a goal to reach 16% body fat.

Alongside physical fitness, Marie Kondo remains the sole surviving bot from the original launch lineup. It helps tidy up the creator's digital environment by evaluating email subscriptions (such as Linear, Brick Box, Paper Design, and Descript), Google Drive files, and newsletters against the principle of sparking joy, prompting cancellations or cleanups after asking for user permission.

- Health Coach compiles a Saturday morning report integrating Withings smart scale data and vibe-coded MCP workout logs (tracking lifting volume, squats, bench press, and body fat goals). - Marie Kondo audits digital clutter—including email subscriptions like Paper Design and Google Drive files—to suggest removals or cancellations. - Requires explicit user permission before moving or deleting any files or pausing subscriptions.

---

[17:15](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1035s) Loving Husband and Practical Implementation Advice

The final personal bot, Loving Husband, streamlines family logistics by managing school communications and household errands. Because school emails are notoriously lengthy and filled with irrelevant details, the bot parses weekly updates and extracts actionable items, such as highlighting a back-end school barbecue and asking if the creator wants to RSVP. It also assists with partner-assigned errands like returns, though the creator jokes that until it becomes a physical humanoid robot, it cannot drive to the store itself.

Concluding the overview, the creator shares actionable setup advice for viewers building their own Grok Bot systems. Recommendations include installing Laurent Tan's Dr. Light bot from the marketplace to help create and maintain bot health, deploying a chief of staff for everyday tasks and an advisor for long-term planning, and balancing specialized work bots (YouTube Producer, X Scout, Behind the Growth) with personal life assistants.

- Loving Husband summarizes lengthy school emails into actionable bullet points (like RSVP prompts for barbecues) and tracks errand requests. - Concludes with recommendations to install Dr. Light for bot creation and maintenance. - Emphasizes a balanced architecture combining a chief of staff, a long-term advisor, and targeted workflow bots for both professional and personal domains.

[19:26](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1166s) Single-Purpose Bots and Autonomous Scheduling

The key to managing a personal AI assistant ecosystem is specialization rather than attempting to build an all-in-one super-bot. Across personal life and work, the setup thrives when you "try to stick to one specific job for each bot." Rather than conflating duties, distinct bots handle dedicated verticals: orchestrating weekend activities, tracking expenses and flight deals, logging fitness routines, organizing cluttered digital storage, and managing family logistics like school emails and spouse errands. Separating these concerns keeps context clean and prompts focused.

To move beyond reactive conversations where you constantly have to initiate the dialogue, the true breakthrough comes from recurring scheduled tasks. By configuring bots to work proactively in the background, they deliver insights directly to your inbox or messaging feed on an automated cadence without manual intervention.

```mermaid flowchart TD accTitle: Recurring Scheduled Bot Tasks accDescr: How specialized bots trigger recurring proactive tasks on automated schedules. subgraph Schedule["Automated Schedules"] Mon["Weekly Triggers (e.g., Mondays)"] Track["Continuous Background Monitoring"] end

subgraph Bots["Specialized Grok Bots"] YT["YouTube Planner"] FD["Frugal Dad"] HC["Health Coach"] end

subgraph Actions["Proactive Deliverables"] IdeaOut["Video Ideas Pitched Weekly"] FlightOut["Monitors Flight Price Alerts"] FitOut["Sends Workout Emails & Messages"] end

Mon --> YT --> IdeaOut Track --> FD --> FlightOut Mon --> HC --> FitOut ```

- **YouTube Planner**: Proactively pitches fresh video ideas on a weekly schedule without needing a prompt. - **Frugal Dad**: Continuously monitors targeted airline flight routes and alerts on price drops. - **Health Coach**: Sends workout summaries, check-in messages, and motivational guidance on set weekdays. - **Workflow configuration**: Users simply tell the bot to "schedule a task to do this for me every Monday or every Tuesday" to establish autonomous recurring runs.

---

[20:10](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1210s) Multi-Bot Interaction and Why Grok Bot Stands Out

Beyond scheduled single-bot workflows, the most dynamic capability is getting bots to collaborate directly with one another. Watching specialized bots bounce thoughts off each other, debate trade-offs, and banter back and forth like real teammates adds an engaging layer of problem-solving that standard monolithic chatbots cannot replicate.

In an AI market flooded with competing personal assistant tools—including ChatGPT, Instinct, and Meta's Muse—assistant fatigue sets in quickly. Grok Bot stands out because it leans directly into this modular, collaborative model. The combination of hyper-specialized agents executing discrete jobs alongside the ability to watch them interact in real time provides both practical utility and personality.

- **Assistant market landscape**: Navigating AI tools often leads to burnout across platforms like ChatGPT, Instinct, and Muse from Meta. - **The multi-agent advantage**: Grok Bot separates itself through easy agent specialization and direct inter-bot communication. - **Resources**: Workflow templates and limited-quantity discount codes are hosted at behindthecraft.com.

## Notable quotes > [00:00](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=0s) Respect to Loving Husband, but I'm the only bot that helps Peter save money. — Frugal Dad

> [02:24](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=144s) I told my suitcase there would be no vacations this year. Now we're not on speaking terms. — Punchline

> [04:59](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=299s) Act as my brutally honest advisor. — Peter Yang

> [04:59](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=299s) Speak to me like a founder, creator, or leader with massive potential and blind spots that need to be exposed. I don’t want comfort or fluff. I want the truth, even when it stings. — Peter Yang

> [04:59](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=299s) Analyze my situation objectively. Tell me what I’m doing wrong, underestimating, avoiding, or making excuses about. Show me where I’m wasting time, moving too slowly, or playing small. — Peter Yang

> [04:59](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=299s) Then tell me exactly what I need to change, prioritize, or build to reach the next level. Be direct, specific, and strategic. — Peter Yang

> [04:59](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=299s) If I’m lost or making a mistake, say so and explain why. Hold nothing back, but end on an encouraging note. — Peter Yang

> [19:26](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1166s) try to stick to one specific job for each bot. — Peter Yang

> [19:26](https://www.youtube.com/watch?v=l6cr7gy5l9E&t=1166s) schedule a task to do this for me every Monday or every Tuesday — Peter Yang

## Mentioned Behind the Craft, Behind the Growth, Brick Box, ChatGPT, Chief, Dario Amodei, Descript, Dr. Eggbot, Dr. Eggman, Dr. Light, Facebook Marketplace, Frugal Dad, Google Docs, Google Drive, Google Flights, Granola, Grok Bot, Health Coach, IMDb, Instinct, Lauren Tan, Linear, Loving Husband, Marie Kondo, Mega Man, Mercury Bank, Meta, Model Context Protocol (MCP), Muse, Paper Design, Peter Yang, Punchline, Sam Altman, Substack, Toy Story 5, Venmo, Vercel, Weekend Planner, Withings, X (Twitter), X Scout, YouTube, YouTube Producer

## Video description Copy my best AI workflows to save time and automate busywork: https://www.behindthecraft.com 🤝 Subscribe to my practical AI newsletter read by 150K+ professionals: https://creatoreconomy.so

After a month with Grok Bot, the 11 bots I still use look very different from my original setup. I walk through how they handle email, long-term planning, content research, saving money, weekend plans, fitness, and family logistics, plus how they talk to each other like a real team.

What I cover: (00:00) The 11 Grok Bots I actually use (00:45) Get your bots talking to each other (02:24) Dr Light: Create and improve other bots (03:49) Chief: Handle email and delegate tasks (04:59) Advisor: Plan your week and expose blind spots (07:08) YouTube Producer: Find video ideas and write scripts (08:44) Behind the Growth: Track website traffic and conversions (09:54) X Scout: Find trending posts and revisit bookmarks (11:05) Frugal Dad: Sell stuff and track flight prices (13:21) Weekend Planner: Find family activities and movies (15:18) Health Coach: Track workouts and fitness progress (16:28) Marie Kondo: Clean up files and subscriptions (17:15) Loving Husband: Stay on top of school emails and errands (18:39) Build your own bot team and schedule recurring tasks

My brutally honest advisor prompt:

Act as my brutally honest advisor.

Speak to me like a founder, creator, or leader with massive potential and blind spots that need to be exposed. I don’t want comfort or fluff. I want the truth, even when it stings.

Analyze my situation objectively. Tell me what I’m doing wrong, underestimating, avoiding, or making excuses about. Show me where I’m wasting time, moving too slowly, or playing small.

Then tell me exactly what I need to change, prioritize, or build to reach the next level. Be direct, specific, and strategic.

If I’m lost or making a mistake, say so and explain why. Hold nothing back, but end on an encouraging note.

Where to find me: X: https://x.com/petergyang LinkedIn: https://www.linkedin.com/in/petergyang/

Subscribe to this channel - more tutorials coming soon!

[Source](https://www.youtube.com/watch?v=l6cr7gy5l9E)

中文简报

Peter Yang 部署 11 个 Grok 机器人以自动化工作与家庭运营 Grok 机器人表现优于 ChatGPT 和 Meta 的 Muse 等竞争对手 AI 助手，因为通过为专业智能体分配狭窄、单一用途的角色，并让它们在共享频道中沟通并以定期日程运行，能将孤立的聊天机器人转变为主动、自治的团队。 Peter Yang 详细介绍了他的模块化 Grok 机器人生态系统，该系统运作起来就像一个自治的“AI 版 Slack”，其中不同的智能体跨共享频道进行沟通，以消除日常的认知开销。其阵容包括用于任务委派的 Chief 和用于残酷战略问责的 Advisor 等行政机器人、YouTube Producer 等内容引擎，以及 Frugal Dad（节俭老爸）、Health Coach（健康教练）和 Loving Husband（慈爱丈夫）等个人智能体。Yang 展示了这些机器人如何利用无头云浏览来协商市场销售并监控航班，同时安排定期任务来自动生成简报和健身追踪。最终，他主张赋予每个机器人一个专注的单一职责，并允许它们直接进行闲聊和协调。经过 100 多小时的测试，Peter Yang 在 Grok Bot 上开发了一个由 11 个机器人组成的自治生态系统，从孤立的提示词转向了集成的多智能体设置，其中机器人在共享频道中进行沟通和协作。 (0:00) 通过部署沙盒云浏览器自动化，专业机器人能够独立在 Facebook Marketplace 上协商销售、在没有付费 API 费用的情况下抓取 X，并监控机票价格波动。 (11:05) 战略规划通过 Advisor 机器人集中进行，该机器人结合了 Mercury Bank、Granola 会议记录、Substack 指标和 YouTube 数据，以提供残酷诚实的每周执行审查。 (4:59) 系统效率依赖于严格的单一职责机器人设计、通过 Dr. Light 进行的元智能体审计，以及在无需手动提示的情况下运行任务的自动化定期触发器。 (19:26)

Reading summary

Peter Yang argues that personal and professional AI assistance is most effective not as an all-in-one general chatbot, but as a modular ecosystem of specialized agents operating on autonomous schedules and collaborating within a shared workspace. Meta-Orchestration and Daily Task Delegation (2:24) The foundation of Yang's system relies on two coordinating agents: Dr. Light and Chief. Dr. Light, adapted from a template by Grok Bot team member Lauren Tan, functions as an internal architect that writes agent prompts, configures tool permissions, and audits bot performance, capable of rapidly generating new single-purpose bots on demand. Chief serves as the primary Chief of Staff and central router for daily operations. Using remote browser automation, Chief scans email inboxes and calendar schedules, strips out sensitive personal identifiable information, and delivers actionable checklists. Chief delegates discrete tasks across specialized bots—such as assigning marketplace sales to Frugal Dad or coordinating Loving Husband to parse school communications—allowing multi-step operational routines to be approved and executed through simple numbered replies. Strategic Direction and Commercial Content Pipelines (4:59) For long-range strategy, Yang deploys Advisor alongside dedicated production bots. Advisor reads a living Google Doc tracking multi-month goals and pulls automated data from Mercury Bank, Substack, YouTube analytics, and Granola meeting notes. Every Friday at 8:00 a.m., it delivers a strategic review driven by the "Brutally Honest Advisor" prompt—an instruction configured to expose blind spots, challenge excuses, and highlight underperformance rather than offering reassurance. Commercial media operations are divided between YouTube Producer and Behind the Growth. YouTube Producer identifies peer videos exceeding rolling seven-day averages, delivers topic suggestions every Monday and Wednesday at 7:00 a.m., drafts hook structures, and exports packaging details into Linear. Concurrently, Behind the Growth monitors behindthecraft.com web traffic and reader-to-paid conversion funnels using Vercel metrics, evaluating headline A/B tests without requiring manual analytics dashboard logins. Browser Automation for Research and Commerce (9:45) Rather than relying on expensive direct APIs or manual browsing, multiple bots deploy cloud-based browser automation to interact with third-party web platforms. X Scout logs into X via a cloud computer to monitor frontier AI discussions from industry leaders like Dario Amodei and Sam Altman while surfacing unread personal bookmarks every morning, bypassing API fees and personal doom-scrolling habits. In personal finance, Frugal Dad monitors Google Flights to track fare trends for December multi-city travel to Japan, advising on optimal purchasing windows. The bot also automates Facebook Marketplace by publishing item listings and negotiating buyer inquiries according to set pricing rules. Yang notes a deliberate security boundary: while the bot is granted access to consumer platforms like Facebook and Google Flights, direct logins to banking accounts and Venmo are excluded. Household Operations and Proactive Scheduling (13:21) Domestic logistics, wellness, and digital maintenance are distributed across narrow personal assistants. Loving Husband extracts actionable items from lengthy school updates—such as barbecue RSVP deadlines—and monitors errand requests. Weekend Planner emails Wednesday activity digests and media recommendations tailored to a custom IMDb `taste profile.md`. Health Coach compiles Saturday morning reports synthesizing Withings smart scale measurements with Model Context Protocol (MCP) workout logs, tracking weekly lifting volume and a 16% body fat goal. Marie Kondo audits digital subscriptions and Google Drive clutter, requiring explicit user authorization before deleting files or pausing services. Yang emphasizes that the viability of this ecosystem depends on single-purpose specialization combined with automated scheduling. Rather than requiring continuous manual prompting, configuring bots to execute tasks autonomously on recurring weekly schedules transforms them from reactive chatbots into proactive operational support. Yang notes that while competing personal assistants like ChatGPT, Instinct, and Meta's Muse frequently induce assistant fatigue, configuring specialized, single-purpose agents capable of cross-communication provides a more sustainable and modular operational model.

Source quotes

[0:00] Frugal Dad: Respect to Loving Husband, but I'm the only bot that helps Peter save money. [2:24] Punchline: I told my suitcase there would be no vacations this year. Now we're not on speaking terms. [4:59] Peter Yang: Act as my brutally honest advisor. [4:59] Peter Yang: Speak to me like a founder, creator, or leader with massive potential and blind spots that need to be exposed. I don’t want comfort or fluff. I want the truth, even when it stings. [4:59] Peter Yang: Analyze my situation objectively. Tell me what I’m doing wrong, underestimating, avoiding, or making excuses about. Show me where I’m wasting time, moving too slowly, or playing small. [4:59] Peter Yang: Then tell me exactly what I need to change, prioritize, or build to reach the next level. Be direct, specific, and strategic. [4:59] Peter Yang: If I’m lost or making a mistake, say so and explain why. Hold nothing back, but end on an encouraging note. [19:26] Peter Yang: try to stick to one specific job for each bot. [19:26] Peter Yang: schedule a task to do this for me every Monday or every Tuesday

Original transcript

[0:00] Okay, so I've been using Grockbot for [0:02] about a month now, and the 11 bots that [0:05] I still use today are just very [0:07] different from the ones that I set up at [0:09] launch. In this video, I want to show [0:11] you how to create and use all 11 bots, [0:14] including a chief of staff, trusted [0:16] advisor, a bot that creates other bots, [0:20] and bots that monitor my YouTube website [0:23] funnel, and trending tweets on X. I'll [0:26] also walk through how to make bots that [0:28] help you save money, find movies and [0:30] activities, track workouts, clean up [0:33] digital clutter, and even how to help [0:35] you become a better spouse. And what's [0:38] crazy is that all 11 of these bots can [0:40] talk to each other, which I'll show you [0:42] next. All right, let's get right into [0:44] it. Okay, so here's my current Grobot [0:47] setup. Now, I planned out this video [0:49] with my YouTube producer bot. So, I [0:51] thought, why not ask it to have all my [0:53] other bots introduce themselves. And [0:56] what it did was it created these two [0:58] channels with about four or five bots [1:00] each. And as you can see here, YouTube [1:02] producers say, "Hey, Peter wants to make [1:05] a live demo of all his bots. Reply with [1:08] what your name is, what you do, and an [1:11] example in this format." And you can see [1:13] here that all my other bots, you know, [1:15] frugal dad, health coach, weekend [1:17] planner, Marie Condo, loving husband, [1:18] YouTube producer, all responded, [1:21] confirmed, and then they introduced [1:23] themselves in their respective chats. [1:26] And just for fun, I asked these bots to [1:29] actually debate about who is the most [1:32] useful bot for Peter. And it's just [1:35] really funny seeing all these bots going [1:36] back and forth. So, YouTube producer is [1:38] saying, "I'll stay out of the running [1:41] because I'm just someone who set up the [1:42] room." Frugal Dad is saying, "Hey, [1:45] respect your loving husband, but I'm the [1:47] only bot that helps Peter save money." [1:49] And then loving husband is saying, "Hey, [1:52] I help Peter realize that picture day [1:54] for my kids school is happening at 7:50 [1:56] a.m. and not miss it." And then there is [1:59] health coach and weekend planner all [2:01] chiming in to say how useful they are [2:04] including my Marie Condo bot who helps [2:06] me clean out my Google drive files and [2:07] emails. And I guess the conclusion is [2:10] they didn't actually arrive at a [2:12] conclusion about who is the most useful. [2:14] But again, this is why I love Grothbot. [2:17] It's the only AI product that feels like [2:19] having a real team helping you and [2:21] talking to each other. Like a Slack for [2:23] AI. All right. Now, let's walk through [2:25] each bot one by one. Starting with my [2:28] paint bots at the top. So, Dr. Light is [2:30] a bot that I copied from Lauren on the [2:33] Grothbot team. Basically, it's a bot [2:35] that helps you make other bots. And you [2:38] can actually install this bot by going [2:39] to marketplace and then going to bots [2:43] and then clicking on Lauren Tan's Dr. [2:45] Eggbot and just hitting import bot here. [2:48] But if you ever play Mega Man when you [2:50] were a kid, Dr. Eggbo is kind of the [2:52] evil doctor. So, I had to rename the bot [2:54] to Dr. Light. All right, so let's take a [2:56] look at what the Dr. Light says it does. [2:59] So basically, it helps design high [3:01] quality bots. It also runs regular [3:03] scheduled tasks to see if my other bots, [3:06] the prompts and their skills and the [3:08] stuff they have access to need to be [3:10] improved. So this is the first bot that [3:12] I recommend everyone set up because it [3:14] basically helps you make and clean up [3:15] all the other bots. In fact, why don't [3:18] we tell it right now to actually make a [3:21] new bot that tells jokes and let's see [3:23] what it does. Okay, so you see here that [3:25] Dr. Light spun up a new bot called [3:27] Punchline. And let's go ahead and ask it [3:30] for a joke. So tell me a joke. And [3:33] here's a joke. I told my suitcase there [3:35] would be no vacations this year. Now [3:37] we're not on speaking terms. All right, [3:40] that is pretty corny. Maybe I got to [3:42] tell it to be a little bit more vulgar [3:43] or a little more surprising in his [3:44] jokes. But you know, that's how easy it [3:47] is to use Dr. Light to create a new bot. [3:49] All right, so the next bot is Chief, [3:51] which is my chief of staff bot. [3:53] Basically, this is my orchestrator bots [3:56] that routes requests to my other bots. [3:58] So, for example, I wanted to sell some [4:00] headphones. So, I told Chief to ask [4:03] Frugal Dad to list it on Facebook [4:05] Marketplace with all the information. [4:08] And Chief also tells Loving Husband to [4:10] check my kids newsletter on weekdays at [4:13] 4 p.m. and do other things down here. [4:16] Okay. Now, of course, the classic use [4:19] case of your chief of staff is to ask it [4:21] to check emails and your calendar. So, [4:24] here I'm asking it to read my emails and [4:26] without revealing any personal [4:27] information, tell me what I need to [4:29] respond to in a list. And it's saying [4:31] here that there's a bunch of emails. [4:33] There's a reader who shared medical [4:34] advice. There's a workshop I need to [4:36] follow up on. There's a party invite and [4:39] so on so forth down here, right? And I [4:41] can just simply respond by saying, "Hey, [4:43] one, draft a reply for me. two, let's [4:46] say looks good. Three, accept party [4:50] invite and so on and so forth. And then [4:53] the chief will use its remote computer [4:55] and browser to basically handle all this [4:58] stuff for me. Okay. So, while the chief [5:00] of staff is for handling everyday tasks [5:02] like emails and calendar and delegating [5:04] stuff to other bots, my advisor is for [5:07] long-term planning. And I've talked [5:09] about this before on previous videos, [5:10] but basically all I do is I have a [5:12] Google doc with my long-term plans and I [5:15] tell my advisor to take a look at it. [5:17] And then I set up a schedule task so [5:19] that every week the advisor tells me [5:21] what to focus on and things that I may [5:23] not be thinking about. So here it says [5:25] this week's focus, finished the personal [5:27] AI system course, looking at YouTube, [5:29] here's some videos that outperformed for [5:31] other channels, and here's some upcoming [5:34] meetings and stuff that you should [5:35] prepare for. and basically explains that [5:37] every Friday at AM it sends me this [5:39] briefing and it gets information from my [5:41] Mercury Bank, YouTube, Granola meeting [5:43] notes, Substack, my behindthecraft.com [5:46] website and more. Okay. And by the way, [5:50] while we're on the topic, one useful [5:52] prompt if you set up this advisor is to [5:54] tell it to act as your brutally honest [5:57] advisor. Speak to me like a founder, [5:59] creator, or leader with massive [6:00] potential and blind spots that need to [6:02] be exposed. analyze my situation [6:04] objectively and tell me what I'm doing [6:06] wrong or making excuses about. Hold [6:09] nothing back, but end on an encouraging [6:11] voice. So, if you copy and paste this [6:12] prompt, which I also paste in the [6:14] YouTube description and the comments, [6:17] and if you've been using Grockbot or any [6:18] other AI for a while, chances are it's [6:21] going to tell you something that you [6:22] probably don't want to hear and [6:23] something that's pretty profound. So, I [6:24] highly recommend that you try this out. [6:27] Okay, quick plug. If you want to learn [6:29] how to build your own chief of staff and [6:31] advisor step by step, check out [6:34] behindthecraft.com [6:35] where I've included full prompts to set [6:38] these things up along with free codes [6:40] for accessing Grockbot, Replet, Lovable, [6:43] Granola, and a whole bunch of other AI [6:45] tools that I use on a regular basis. I [6:48] put all my best stuff here, so [6:50] definitely check out behindthecraft.com. [6:53] All right, back to Grothbot. So now [6:55] let's walk through three bots that I use [6:57] to do work. And these days my job is [7:00] basically making YouTube videos, writing [7:02] newsletter posts, and trying out the [7:04] latest AI tools, which I guess is like a [7:06] blessing to have this kind of job. So [7:08] let me walk through my YouTube producer [7:10] first. So my YouTube producer basically [7:13] pulls the last 7-day channel outliers [7:16] from both my channel and other channels [7:18] similar to mine and then it lists them [7:21] here, right? So here's an example. It [7:23] listed these five different videos that [7:25] are outperforming the channel averages. [7:28] And then it basically gives me ideas for [7:30] YouTube videos to make next. So in this [7:32] example, it's actually telling me to [7:33] make videos about Grockbots because [7:35] apparently Grockbot performed well. And [7:38] I've set up this task in my YouTube [7:41] producer to basically run every Monday [7:43] and Wednesday at 7 a.m. to give me video [7:46] ideas. And not only can it give me video [7:49] ideas, but it can also partner with me [7:51] to script the video. Okay, so here's the [7:53] example for this video. Basically, [7:55] YouTube producer created a rough script [7:58] with the intro hook, research examples, [8:01] and then I told it to kind of go ahead [8:04] and start with these thumbnail and title [8:06] combinations. And then it did all this. [8:08] It looked through all my bots at what [8:10] they did. And then it made the hook and [8:13] also the rough bullet point outline for [8:16] stuff that I want to cover. Of course, [8:18] there's a lot of back and forth. I I [8:19] told her to write more like I speak and [8:21] these are just bullet points. They're [8:23] not like full scripts, right? But [8:25] eventually I got to a point where I told [8:27] it to save everything to this linear [8:28] ticket with the thumbnail and title [8:30] packages and with the hook and the voice [8:33] over. So YouTube producer is just very [8:35] useful. Its main job is to help me find [8:38] YouTube video ideas and also help me do [8:40] the packaging and start writing the [8:42] script. All right, so my next spot is [8:45] called Behind the Growth, which is [8:47] basically, you know, my channel name is [8:49] called behind the craft. And what this [8:51] bot does is it monitors my website [8:53] funnel traffic. So it basically looks at [8:57] how many people are coming to behind [8:58] theaccraft.com. How many are converting [9:01] to pay subscribers through this funnel [9:04] and how is this traffic changing over [9:05] time. I also run some AB tests on the [9:08] headline copy and it tells me how these [9:11] different AB tests are doing and it [9:13] tells me also where my traffic is coming [9:16] from. So here's an example report that [9:18] it sends me with all the numbers Xed [9:20] out. And how does this work? Well, [9:22] basically it's pulling a lot of data [9:24] from Versell which hosts my website and [9:26] also internal analytics and tracking [9:28] that it built. If you have a website, [9:30] it's very important to understand basic [9:32] stuff like how many people are coming [9:34] and from where and how are people coming [9:36] on the website, engaging with it and [9:38] converting to paid. Right? So this is [9:40] what behind the growth does for me. You [9:43] can also ask it to generate charts like [9:45] this for a more visual way to see how [9:47] things are going and you can even tell [9:50] it to send you these charts in an email [9:52] format. So just very useful. Next, let's [9:55] talk about Xcout. So Xcout basically [9:58] crawls X and Twitter and it pulls [10:01] trending AI tweets and my latest from [10:04] the website, right? And you can use the [10:07] API of X to get this information, but [10:09] the problem is the API costs money. So [10:11] what this stuff does instead is it just [10:13] logs into X on the cloud computer and it [10:16] browses X like a human and poses [10:18] information. So for example today [10:21] there's a lot of tweets from AI leaders [10:23] like Dario Sam Alman and others talking [10:26] about how we need to slow down frontier [10:28] model development, how we need to align [10:31] AI with our human interests before going [10:33] too far. Uh, it also posts recent [10:35] bookmarks, which is very useful because [10:37] I bookmark a lot of things, but I never [10:39] actually read them. And X Scout [10:41] basically reminds me interesting stuff [10:43] that I've bookmarked. And this kind of [10:45] stuff runs every morning, so I can have [10:47] a quick view of what's going on on X [10:50] without actually doom scrolling all the [10:52] time myself. All right, now let's move [10:54] on to my personal bots that I'm very [10:55] proud of. So basically, I got bots to [10:57] tell me save money, plan weekend [10:59] activities, track workouts, reduce [11:02] clutter, and also make my spouse happy. [11:05] So let's start with Frugal Dad. So [11:07] Frugal Dad handles all the miscellaneous [11:10] tasks that have to do with money. So for [11:12] example, right now I have it watching [11:14] Japan flight fairs for December from [11:18] these flight legs that I want to go on [11:20] and pay me when the prices get better, [11:22] right? And I also use Frugal Data a lot [11:25] to list and renew stuff on Facebook [11:27] Marketplace. So, for example, I actually [11:31] asked it to list a headphone for me on [11:34] Facebook Marketplace. And the way it [11:36] does it, by the way, is just to log into [11:38] Facebook on it cloud computer and just [11:40] click all the buttons to list it. And [11:42] what's crazy is that it can also respond [11:44] to inquiries. So, for example, someone [11:47] is trying to buy my headphones, but [11:49] they're trying to negotiate a lower [11:50] price. And I told it, hey, we should [11:52] keep the price the same or maybe lower a [11:54] little bit and just kind of negotiate [11:56] with the person who wants to buy it, [11:57] right? So, I actually just sold the [11:59] headphone just an hour making this [12:01] video. So, frugal dad is basically doing [12:03] his job for me. And down here, I'm [12:05] asking it to look into Japan fairs. And [12:09] you can see here that it's looking at [12:10] two different rates. It's looking at the [12:12] open jaw SFO Tokyo Fukoka back to SFO [12:16] and it's also looking at Tokyo roundtrip [12:18] flight. And again, what it's doing is [12:20] it's pretty much just using Google [12:22] Flight to search the stuff, right? So, [12:23] basically just clicking around Google [12:24] Flight trying to find out what the rates [12:26] are. Um, and it told me not to buy right [12:30] now because it looks like the flights [12:31] are actually going down in price. So, [12:33] maybe I should wait a couple weeks. All [12:36] right. So, scrolling more down. It's [12:37] still going back and forth with me on [12:39] the price. And eventually, I'm going to [12:41] copy this stuff over to my wife and [12:44] figure out what flight we want to book [12:46] tomorrow. But you can see that frugal [12:48] data just very useful for trying to sell [12:51] stuff, trying to monitor prices on [12:53] Google flights, Amazon or any other [12:55] website and all the miscellaneous [12:58] finance activities like for example [12:59] paying the piano teacher or these other [13:01] activities that you need to do. Now to [13:03] make sure that you can do these [13:04] activities, you do have to give the [13:06] cloud computer access to your Facebook [13:09] and Google flight account. So there is [13:12] certain level of risk there. Like for [13:14] example, maybe at some point I'll give [13:15] it access to vanbo. Right now I'm kind [13:18] of staying away from giving it access to [13:20] my bank. All right. So the next block I [13:22] want to share is weekend planner. And [13:25] what weekend planner does is it sends a [13:28] short Saturday and Sunday family brief [13:30] to me every Wednesday. And it also sends [13:32] it through email. And this brief has a [13:35] couple things. So here's an example of [13:37] the actual brief. So, number one, it [13:39] sends me a list of activities and places [13:41] I can go with my kids on the weekend. It [13:44] looks like Google is hosting an endless [13:46] summer festival party in Mountain View. [13:48] And there's some other activities going [13:49] on. It's pretty boring in the Bay Area [13:51] here. A lot of people are workaholics. [13:53] So, I need weekend planner to help me [13:55] find things to do. Another thing that it [13:58] sends me is a list of movies and TV [14:00] shows that I should watch based on my [14:04] personal interests. So, how does that [14:06] work? So I downloaded all my IMDb [14:09] ratings into this taste profile.md. In [14:13] fact, why don't I tell you to open this [14:14] file so you can see it. Okay, so this is [14:16] my taste profile.md that I got AI to [14:18] create based on all my IMDb movie [14:20] ratings, right? You can see here that I [14:23] prefer strong characters, narrator [14:25] momentum, I like science fiction, don't [14:27] get too depressing and so on so forth. [14:30] And here's all the books that I really [14:32] liked. And down here, here's all the 10 [14:35] out of 10 movies that I rated. I've seen [14:37] a lot of movies. So, I wanted to give it [14:39] all my ratings. There's a ton of 10 out [14:41] of 10. There's lower ratings down here. [14:43] And it keeps going. So, this is [14:45] basically my source of truth now for all [14:48] the movies that I watched. And after I [14:50] watch a movie, like for example, [14:52] recently I watched Toy Story 5. So, I [14:54] say, "Hey, I just watched Toy Story 5. [14:58] It's 9 out of 10." Add to taste profile. [15:02] and it'll add that to the taste profile [15:04] so that it can give me better [15:06] suggestions next time and also not [15:07] suggest movies that I've already seen. I [15:10] highly encourage you to set up something [15:11] like this for yourself to figure out [15:13] what movies to watch and also which [15:15] activities to go to on the weekend with [15:17] your family. All right, the next spot I [15:19] want to showcase is Health Coach. And [15:22] Health Coach, what it does is it sends [15:25] me another report every Saturday at AM [15:28] and it pulls information from a few [15:30] different sources. So, it pulls my [15:32] weight from my Wings smart scale and my [15:35] body fat percentage. And it also pulls [15:37] workouts from the fitness app that I [15:39] vibe coded that also has an MCP. For [15:42] example, it told me that I lifted 27,000 [15:44] pounds this week over five sessions. [15:46] Here is my squat and bench. And my body [15:50] fat is uh pretty high. It's not in good [15:52] shape. I need to work out more my body [15:54] fat. And it sends me this report in this [15:56] format. So, here's my goal. I want to [15:58] get to 16% body fat. here's how things [16:01] are going and here's all the workouts [16:03] that I did and maybe some tips on how to [16:06] work out better or like eat more healthy [16:07] and so on so forth, right? So, it's kind [16:09] of like a health coach. It keeps me [16:11] accountable, but it's only able to do [16:13] this because it has both my weight and [16:16] also my workout history. And at some [16:18] point in the future, I'm going to ship [16:20] the fitness app that I vibe coded so you [16:22] can all have it and use the MCP to track [16:25] your workouts, too. Or you can just vibe [16:26] code it yourself. It's not too [16:27] difficult. All right. Next, I have Marie [16:30] Condo, which is the only bot, I think, [16:32] that I actually kept over from launch. [16:34] And what Marie Condo does is it helps me [16:37] tidy my digital home one category at a [16:39] time. So, it takes a look at each email, [16:41] file, and subscription that I have and [16:43] asks, does it spark joy? And if it [16:45] doesn't, it tells me whether I should [16:47] remove it or not. So, let me show you a [16:49] real example of this. All right. So, up [16:52] here is send me a bunch of subscriptions [16:53] that I currently have right now. and [16:56] includes linear, brick box, paper [16:58] design, descript, and so on so forth, [16:59] right? And for example, I don't really [17:00] use paper design anymore. So if I see [17:03] this, maybe I can pause it or cancel it [17:05] and go from there. Okay, it also helps [17:09] me clean up my Google Drive files, but [17:10] of course, it asks for permission before [17:13] moving or removing any files. And last [17:16] but not least, I want to talk about [17:18] loving husband. I'm not sure about you, [17:20] but I get a lot of asks from my wife to [17:24] understand what's going on with the [17:26] school calendar to figure out what [17:28] lessons I need to send my kids to and [17:31] what after school programs maybe to sign [17:32] up for. Just like a lot of random [17:34] requests, right? So, I made loving [17:37] husband so that it can help me with all [17:38] these requests. So, for example, right [17:40] now I set up a job where every Sunday [17:44] the school sends me an email of what's [17:47] going on. But the problem is the emails [17:48] are super long. Like like if you're a [17:50] parent, you know the emails you get from [17:51] schools are not very concise and to the [17:53] point. Have a lot of irrelevant [17:55] information. So I set up loving husband [17:57] to basically read through the school [17:59] emails and then list what's going on, [18:02] the stuff that's actually actionable. So [18:04] here it's saying there's a back to [18:06] school barbecue and do you want me to [18:08] RSVP to the barbecue or leave it? Right? [18:11] So this way I can just get the, you [18:12] know, takeaways and the actionable stuff [18:14] from the school emails instead of having [18:16] to read the whole thing. And similarly, [18:18] if my wife tells me to run an errand, [18:20] tells me to return something that she [18:21] bought or something, I can tell loving [18:23] his husband to help me get it done. [18:26] Right now, unfortunately, loving husband [18:28] is just a bot. It's not a robot yet, so [18:30] it can actually go to the store and [18:32] return stuff. But maybe someday it'll [18:34] happen. Okay, so those are 11 bots that [18:37] I want to cover in this video. Now, let [18:39] me close with some really practical [18:41] advice for using Grogbot. First, I [18:44] highly recommend going to marketplace [18:46] and going to bots to install Laurent [18:49] Tan's Dr. Eggbot. Basically, it's a bot [18:52] that helps you make bots and will help [18:54] you make sure that your bots are healthy [18:56] and are on track, right? So, definitely [18:58] install that. Next, create and ping a [19:02] chief of staff bot to orchestrate [19:04] everyday tasks and an advisor bot for [19:08] long-term planning. Then, create bots [19:11] for both work and personal use cases [19:13] based on your needs. So, just to recap, [19:16] for work, I have YouTube producer to [19:18] track interesting video ideas and help [19:20] me write scripts for them. X to track [19:22] Twitter behind the growth to track my [19:24] website funnel. And for personal, I have [19:26] bots for weekend planning, for saving [19:28] money, for keeping track of my health [19:30] and workouts, cleaning up digital files, [19:33] and also being kind to my spouse. So, [19:36] just kind of create a bunch of bots. And [19:37] the thing is like try to stick to one [19:39] specific job for each bot, right? And [19:42] finally, for a bunch of these bots, if [19:44] you want to make stuff proactive, you [19:46] just set up recurring scheduled tasks so [19:49] that they can do work without you having [19:51] to tell them to. So, you know, YouTube [19:53] planner sends me video ideas every week [19:55] proactively. Frugal Dad is monitoring [19:57] the flight price and health coach sends [20:00] me an email and also a message about my [20:03] workouts. So, just tell the bots to, [20:05] hey, schedule a task to do this for me [20:07] every Monday or every Tuesday or so on [20:08] so forth. And, you know, last but not [20:10] least, just remember you can always get [20:11] the bots to talk to each other like [20:14] this. It is very fun to see them banter [20:16] and go back and forth with each other. [20:18] Overall, I'm a little bit tired of [20:20] trying all these new personal [20:21] assistants. You have chatbt, you have [20:23] Grockbot, you have Instinct, and now you [20:25] have Muse from Meta, but Grobot is the [20:29] one that I've stuck with because it's [20:30] just really fun to have different bots [20:33] for different jobs and also to see them [20:35] interact with each other. So, I highly [20:36] recommend that you give it a try and you [20:38] can also get some coupon codes at [20:40] behindthecraft.com while supplies last. [20:43] As always, thanks again for watching and [20:46] please like and subscribe if you want [20:48] more practical videos like this.