7 passos obrigatórios para usar o Hermes e criar agentes que trabalham por você (guia completo)
Introduction to Hermes and Its Potential
Overview of the Video
- The speaker, Bruno Camoto, introduces a comprehensive masterclass on using Hermes, an AI tool, from an entrepreneurial perspective. He emphasizes its practical applications in generating value for businesses and teams.
- Bruno mentions that he has invested over 100 hours into exploring Hermes to help viewers save time and money while enhancing their skills with the tool.
Learning Objectives
- The video aims to cover everything from basic to advanced functionalities of Hermes as of May 2026, making it one of the most complete resources available on AI agents.
Background of Bruno Camoto
Entrepreneurial Journey
- Bruno shares his background in technology and entrepreneurship, highlighting his experiences with startups, including successes and failures. He has built a community of over 20,000 entrepreneurs focused on micro-SaaS solutions.
- He is currently working on creating agent-based structures for companies through his new venture called Pixel AI Hub or Pixel Educação. This video serves as a practical guide based on his real-world use of Hermes in business operations.
Contextualizing Hermes
Use Case Example
- Bruno references a previous video where he scaled his startup's revenue significantly within 45 days using OpenAI tools alongside multiple agents across various functions within his team. This context sets the stage for discussing how Hermes fits into similar operational frameworks.
Introducing Dexter: The Operational Agent
What is Dexter?
- Dexter is introduced as an operational layer created using Hermes rather than just a chatbot; it automates content production within Bruno's company and accounts for over 40% of sales. This highlights its importance in business strategy.
Functionality Overview
- Dexter operates through channels like "trending," which aggregates popular content from platforms such as Twitter and YouTube using APIs for data scraping purposes (e.g., Grock Premium). This allows him to stay updated with current trends effectively.
Content Management System via Dexter
Inbox System
- The inbox system categorizes incoming content based on relevance scores (up to 100 points), analyzing factors like timing and fit with existing strategies before sorting them into "ideas" or "hot ideas." This structured approach enhances decision-making regarding content creation.
News Aggregation Channel
AI News Channel
- An additional channel called "AI News" helps track industry news by following creators who specialize in AI updates, providing insights into market trends relevant to the business landscape that can inform strategic decisions moving forward.
Practical Application Examples
Scoring System Insights
- Each piece of content is evaluated against specific criteria before being categorized; this includes assessing its potential impact based on trending metrics across various platforms (e.g., GitHub discussions). Such evaluations ensure high-quality output aligned with audience interests.
Governance in Using AI Tools
Importance of Governance
- A key takeaway is that effective governance ensures that AI tools like Dexter work optimally; without proper oversight, they may become liabilities rather than assets by failing to deliver meaningful results or insights relevant to business goals. Thus emphasizing responsible usage practices when implementing these technologies within organizations.
Enhancements Through Cloud Code Integration
Technical Improvements
- Bruno discusses integrating Cloud Code into Dexter’s framework for improved presentation capabilities during video production processes—demonstrating how technical enhancements can streamline workflows further while maintaining quality outputs across different media formats.
Understanding the Architecture Behind Hermes
Structural Components
- The architecture consists of several layers: memory management systems designed specifically for task execution efficiency along with skill integration allowing seamless interaction between human operators and automated agents—highlighting how well-designed infrastructures can enhance productivity levels significantly when utilized correctly.
Recommendations for Implementing Agents
Best Practices
- For those looking to implement agents like Hermes effectively:
- Start by having your agent read existing structures before revising them.
- Gradually introduce automation features once foundational understanding has been established.
- Focus on building robust systems first so that any automation applied does not exacerbate existing inefficiencies but instead improves overall performance metrics across departments involved.
This structured summary captures essential insights from the transcript while linking back directly to specific timestamps for easy reference during review sessions or further exploration into particular topics discussed throughout the video presentation by Bruno Camoto regarding utilizing advanced AI tools such as Hermes effectively within entrepreneurial contexts today!
Understanding the Concept of a "Second Brain" in Business
The Corporate Second Brain
- The corporate second brain, referred to as Pixel Educação, is where all content, support tickets, decisions, and data are stored.
- Léo, an agent responsible for auditing this system daily, ensures that all information is organized effectively.
Personal Second Brain
- The personal second brain is represented by Amora, which operates across multiple platforms including cloud services and VPS.
- Amora's architecture includes memory components such as context, projects, sessions, integrations, and tracking.
Tools for Building a Second Brain
- Both GitHub and Obsidian can be used to create a second brain; however, GitHub is more suited for agents while Obsidian caters to human users.
- Understanding the structure of a second brain involves recognizing three key elements: context (data), skills (repetitive processes), and routine (automations).
The Importance of Context in AI Operations
Defining Context
- Context consists of data and information necessary for agents to perform tasks effectively; it includes meeting transcriptions and communication logs.
- All generated content from Dexter is saved within an inbox system that organizes inputs from various agents.
Auditing Process
- Léo audits the inbox regularly to categorize information appropriately; if uncertain about categorization, he consults team members.
Creating Structured Second Brains
Structuring Access Levels
- Different access levels can be established within the second brain framework based on roles or departments (e.g., marketing or logistics).
Example of Cloud Integration
- A skill called "brain" executes commands by reading documents across various tools like Notion and emails to maintain organization.
Operational Efficiency Through Automation
Real-Time Collaboration
- Agents read incoming data continuously; this allows real-time collaboration on tasks across different areas of the business.
Universal Contextual Framework
- The concept of a universal context enables seamless brainstorming and task execution without losing track of ongoing projects.
Introducing Pixel A Hub for Entrepreneurs
Educational Resource
- Pixel A Hub offers resources on creating agents and implementing a second brain in businesses with over 500 students enrolled shortly after launch.
Marketing OS Development
- New features like Marketing OS will enhance marketing capabilities through creative content generation.
Exploring Hermes Memory Architecture
Four-Layer Memory Structure
- Hermes utilizes a four-layer memory architecture that provides competitive advantages over other systems like OpenCall.
Peer Cards Feature
- Peer cards store individual user preferences and historical interactions with Hermes for personalized responses.
Fuz Index Functionality
- This feature functions similarly to semantic memory by allowing keyword searches within stored contexts efficiently.
Memory Management Techniques in Hermes
Worker Memory Review System
- Hermes employs an automated worker memory review process that decides what information should be retained after conversations.
Compacting Strategies
- Users can adjust settings related to how aggressively Hermes compacts its memory storage based on their needs.
Recommendations for Optimal Use
- It’s advised to set compacting parameters around 0.8 to avoid overflow issues while maintaining efficient performance.
Skills Development Within Hermes
Automatic Skill Creation
- Every action taken within Hermes automatically generates new skills based on user interactions without manual input required.
Long-Term Maintenance Concerns
- As skills accumulate over time (potentially thousands), there may be challenges regarding maintenance efficiency affecting overall performance.
Introduction to Hermes and Its Importance
Overview of Hermes
- The speaker introduces the concept of "sou" (identity) in relation to agents, emphasizing the need for customization based on individual structures.
- Highlights the significance of combining memory structure with a second brain, suggesting it enhances operational efficiency.
Gateway to Hermes
- Introduction to the "gateway do Hermes," where the speaker demonstrates accessing the Hermes dashboard via SSH.
- The dashboard is described as both exciting and confusing, but ultimately superior to previous tools like doen cl.
Installing and Updating Hermes
Installation Tips
- Advises users to keep their Hermes updated due to its precise updates compared to others that may release more frequently.
- Discusses connecting various sessions through platforms like Telegram and WhatsApp, showcasing recent activity logs.
Features of Hermes
- Users can customize models for specific tasks, enhancing flexibility in content creation processes.
- Mentions available logs for error tracking and managing scheduled tasks within the platform.
Exploring Skills and Plugins
Skill Management
- Describes various skills included with Hermes, some being entertaining yet not necessarily useful for business operations.
- Lists plugins available for installation from GitHub repositories, allowing further customization of user experience.
User Profiles and Settings
- Users can create profiles for different agents or individuals utilizing Hermes, enhancing personalized experiences.
- Configuration options allow adjustments in agent settings including display layout preferences.
Operational Setup with VPS
Recommended Practices
- Emphasizes using a VPS over local installations for security reasons; ensures continuous operation without interruptions from power outages.
Step-by-Step Installation Process
- The speaker shares personal experience creating an account on Hostinger while guiding viewers through selecting plans and applying discounts.
Connecting Agents and Managing Operations
Using Tail Scale
- Introduces Tail Scale as a tool for managing multiple agents securely across VPS environments.
MCP Functionality
- Explains how MCP allows management of agents within Hostinger's environment, including backup capabilities.
Finalizing Agent Setup
Troubleshooting Common Issues
- Addresses potential errors during setup related to OpenAI integration; encourages users not to panic if they encounter similar issues.
Connecting Telegram Bot
- Provides instructions on creating a Telegram bot using Bot Father; emphasizes copying tokens correctly for successful integration.
Best Practices for Agent Management
Treating Agents as Employees
- Recommends treating each agent as an employee by assigning them unique emails and controlled access levels within systems.
Security Measures
- Advises against storing sensitive information directly in agents; suggests using password managers like One Password or Bitwarden instead.
Governance Structure
- Proposes having a dedicated governance agent overseeing all other agents' activities ensuring smooth operations without risks associated with autonomy.
Exploring Heart Beat and LLM Recommendations
Recommendations for Learning about Heart Beat
- The speaker suggests watching a YouTube video on Heart Beat for further learning.
- Recommends using Chat GPT 5.5 due to its high performance, while also encouraging experimentation with other models like Oppos and Haiko for specific tasks.
Utilizing Cloud Code
- Mentions the installation of Cloud Code CLI in a dedicated brainstorming channel for enhanced functionality.
- Emphasizes that without the right tools, agents can only provide basic responses.
Tools and Resources
- Lists several useful tools: Perpecity, Firecrawl (which converts websites into Markdown), APIFI (for YouTube content research), Grock, GitHub (as a second brain), Telegram, Chrome Jobs, Cloud Code Celai, Open Claw, and Scrappers.
- Discusses using Open Claw to audit Dexter as part of governance agent activities.
Key Commands in Hermes
Important Commands Overview
- Introduces the "go" command which allows users to set objectives that Hermes will work on until completion.
- Highlights the importance of specificity in commands; more detailed instructions yield better results.
Interactive Features of Hermes
- Describes how users can interact with Hermes while it is performing tasks by asking questions that interrupt its current activity.
- Introduces the "background" feature allowing users to ask questions while Hermes continues working on other tasks simultaneously.
Additional Functionalities and Tips
Exploring More Features
- Mentions additional commands like "steer" for injecting tools and "curator" for maintaining skills.
- Discusses the summarization feature where users can pull summaries from ongoing tasks or previous work.
Final Advice on Agent Creation
- Advises starting with one useful agent rather than creating multiple agents at once; emphasizes building a solid foundation first.
Closing Thoughts on Tool Utilization
Importance of Implementation Over Comparison
- Encourages viewers not to focus excessively on comparing different tools but rather to implement them effectively in their workflows.
Engagement Invitation
- Invites viewers to share their thoughts about Hermes in comments and encourages likes and subscriptions for more content.