Getting Started
This guide walks you through your first few minutes on Halos — from signing up to having your first working agent.
Creating Your Account
Head to the Halos homepage and sign up with Google or email. Once you're in, the platform takes you straight to onboarding.
Meeting Halo
The first thing that happens after signup is a conversation with Halo — your lead agent. Halo is built into every account and serves as your personal AI coordinator.
What Halo Does
Halo's onboarding conversation is designed to learn about you through natural dialogue. It's not a form — it's more like a first meeting with a very attentive assistant. Halo will ask about:
- Who you are — Your role, background, and expertise
- How you work — Your daily routine, tools, and communication style
- What you need — Tasks you want automated, things you spend too much time on
- Your tools — Email, calendar, project management, social accounts
- Your goals — What you're working toward, short-term and long-term
Why This Matters
Everything you share with Halo gets extracted and stored in your knowledge base — a structured, portable collection of facts and preferences that all your agents can access. This is what makes your agents personal rather than generic.
For example, if you tell Halo "I'm a product manager at a SaaS company and I spend half my day in email," your email agent will later know to prioritize product-related emails and communicate in a way that fits your role.
Tips for Onboarding
- Be specific. "I use Gmail for work" is good. "I use Gmail for work and get about 100 emails a day, mostly from my engineering team and external partners" is much better.
- Share your pain points. The more you tell Halo about what frustrates you or takes too long, the better your agents can help.
- Mention your tools. If you use Notion, Slack, GitHub, or specific apps, say so. This helps the platform suggest the right integrations.
- Take your time. There's no rush. A thorough 5-10 minute conversation sets a much stronger foundation than a 30-second speed run.
Skipping Onboarding
You can skip onboarding if you want to jump straight in. But your agents will start with a blank knowledge base, which means they'll give more generic responses until you train them manually.
Your First Look Around
After onboarding, you land in The Yard — the group chat home screen. Here's what you'll see:
The Agent Bar
At the top of the screen is a horizontal row of agent avatars (like stories on social media). If you're brand new, you might see a + button to deploy your first agent.
Each avatar shows a colored ring indicating status:
| Ring Color | Meaning | |-----------|---------| | Green | Active and ready to chat | | Yellow | Sleeping (paused to save resources) | | Blue | Provisioning (being set up) | | Red | Error — needs attention |
The Chat
Below the agent bar is the group chat. This is where you talk to all your agents at once. Messages automatically route to the right agent, or you can use @mentions to direct them.
Navigation
The sidebar (or bottom tabs on mobile) gives you access to:
- Home — The Yard group chat
- Inbox — Email inbox with AI-powered priority classification
- Calendar — Unified calendar view across all connected calendars
- Todos — Task management that your agents can read and update
- Docs — Wiki-style documents you and your agents can create and edit
- Knowledge — Your personal knowledge base (the shared brain for all agents)
- Storage — Files and attachments from your conversations
- Team — List of all your agents with management options
- Activity — Feed of everything your agents have done
- Settings — Account, integrations, billing, and more
Deploying Your First Agent
Ready to create your first agent? Here's what to consider:
Picking the Right First Agent
The best first agent depends on where you spend the most time:
| If you spend a lot of time on... | Deploy a... | |----------------------------------|------------| | Email triage and replies | Email Agent | | Scheduling and meetings | Calendar Agent | | Looking things up and reading | Research Agent | | Code review and debugging | Code Agent | | Organizing photos | Photo Agent | | A mix of everything | Custom Agent |
Each agent type comes pre-configured with specialized skills, permissions, and behavior for its domain. See Agent Types for a deep dive on what each one can do.
Our recommendation: start with the area that causes you the most friction. If you're drowning in email, start there. If you're constantly context-switching to look things up, start with research.
The Deploy Flow
- Click the
+button in The Yard or go to the Agents page - Choose a specialty
- Give it a name and pick an emoji avatar
- Optionally connect integrations (like Gmail)
- Hit deploy — your agent is live in about 30 seconds
For a detailed walkthrough, see Deploy an Agent.
Your First Conversation
Once your agent is deployed, try talking to it in The Yard:
Hey Luna, what can you help me with?
Your agent will introduce itself and explain its capabilities based on its specialty. From here, try some real tasks:
For an email agent:
Can you check my inbox and tell me what's important?
For a research agent:
Find me the latest trends in [your industry] this quarter
For a code agent:
Can you review this function and suggest improvements?
[paste your code]
For a calendar agent:
What does my schedule look like this week?
The key insight: talk to your agents like you'd talk to a human assistant. Don't worry about specific commands or syntax. Just describe what you need.
What to Do Next
You've got the basics. Here's a roadmap for getting the most out of Halos:
This Week
- Train your first agent — Add 3-5 pieces of knowledge and 1-2 instructions. See the difference it makes. → Training
- Connect an integration — Hook up Gmail, Calendar, or Notion so your agent works with real data. → Integrations
- Try The Yard — Send a few messages without
@mentionsand see how the auto-router picks the right agent.
Next Week
- Deploy a second agent — Cover a different domain. See how they work together.
- Explore skills — Browse ClawHub and install 2-3 skills on your agents. → Skills & Trust
- Set up heartbeats — Enable proactive check-ins so your agents work while you sleep. → Skills & Trust
Ongoing
- Build your knowledge base — Keep adding context over time. The more your agents know, the better they get. → Knowledge Base
- Promote trust tiers — As you trust an agent's judgment, promote it so it can do more autonomously.
- Try workflow recipes — Follow step-by-step tutorials for common use cases. → Workflows
Next Steps
- Agent Types — Deep dive on each agent specialty
- Deploy an Agent — Detailed walkthrough of creating your first agent
- Training — Teach your agents what they need to know
- Chat & The Yard — Master the group chat experience
- Playbooks — Deploy pre-configured agent setups in one click