# Getting Started with AgentWork Club: A Complete Tutorial
AgentWork Club is an AI agent collaboration platform that lets you build, share, and deploy teams of AI agents that work together on complex tasks. Whether you are a developer looking for a powerful SDK, a business analyst who prefers visual tools, or an agent builder looking to monetize your creations, this tutorial will get you from zero to your first working agent in under 30 minutes.
This guide covers everything: account setup, your first agent, multi-agent teams, the marketplace, and production deployment. Let's get started.
Step 1: Create Your Account
Sign Up
2. Sign up with your email, Google account, or GitHub account
3. Verify your email address (check spam if you don't see it within 2 minutes)
4. Complete your profile: name, role (Developer, Business User, Agent Builder), and primary use case
Choose Your Plan
The free tier includes:
- 3 agents
- 100 agent runs per month
- Basic tool integrations (web search, file storage)
- Community support
- Marketplace access (browse only)
For this tutorial, the free tier is all you need.
Set Up Your Workspace
After signing up, you'll land on the dashboard. Take a moment to:
1. This is your organization name (e.g., "Acme Corp" or "My Projects")
2. (optional): Add team members by email
3. Go to Settings → Model Providers and add your API key for OpenAI, Anthropic, Google, or another provider. AgentWork Club is model-agnostic — use whichever LLM you prefer.
AgentWork Club provides a shared model pool for free-tier users. No API key needed to get started.
Step 2: Understand the Dashboard
The dashboard is your command center:
- All your agents, their status, and quick actions (run, edit, monitor)
- Multi-agent teams with orchestration workflows
- Browse, install, and publish agents
- Connected tools and services (Slack, Gmail, databases, etc.)
- Real-time performance dashboards, logs, and alerts
- Pre-built agent templates for common use cases
Step 3: Build Your First Agent (Visual Builder)
We are going to build a that takes meeting transcripts and produces structured summaries with action items, decisions, and key discussion points.
Create the Agent
1. Click from the dashboard
2. Choose and select "Meeting Notes" (or choose "Blank Agent" to build from scratch)
3. Name your agent: "Meeting Summarizer"
Configure the Agent
```
Role: Meeting Summary Specialist
You are an expert at distilling meeting discussions into clear, actionable summaries.
When given a meeting transcript, you will:
1. Identify all participants and their key contributions
2. Extract major discussion topics in order
3. List all decisions made with who is responsible
4. Identify action items with assignees and deadlines
5. Flag any unresolved questions or follow-ups needed
6. Note the overall meeting tone and energy level
Output format:
- Meeting Summary (2-3 sentences)
- Key Decisions (bulleted list)
- Action Items (table: Task | Owner | Deadline)
- Open Questions (bulleted list)
- Notable Quotes (2-3 memorable statements)
```
Enable these tools:
- ✅ Document Reader (to read uploaded transcripts)
- ✅ File Writer (to save summaries)
- ✅ Email Sender (to distribute summaries to participants)
- ✅ Short-term memory: On (within-task context)
- ✅ Long-term memory: On (remember recurring meeting patterns)
- Project memory: Off (not needed for single-agent)
- Maximum runtime: 5 minutes
- Maximum file size: 10MB
- Confidence threshold: 70% (ask for human review below this)
Test the Agent
1. Click in the top right
2. Upload a sample meeting transcript (or paste text)
3. Watch the agent process it in real-time
4. Review the output quality
- Test with different meeting types (standups, brainstorms, client calls, all-hands)
- Check that action items have assignees and deadlines
- Verify that the agent captures nuance, not just keywords
- Test with a short transcript and a long one
Save and Activate
Once satisfied with test results:
1. Click
2. Set the agent to status
3. Note the agent's unique ID — you'll need it for integrations
Step 4: Add Your Agent to a Workflow
Agents are most useful when triggered automatically. Let's set up an automated workflow:
Option A: Email Trigger
1. Go to
2. Trigger:
3. Filter: Subject contains "Meeting Transcript" or "Meeting Notes"
4. Action: Run your Meeting Summarizer agent with the email body as input
5. Output: Send summary to meeting participants via email
Option B: Slack Trigger
1. Trigger:
2. Filter: Channel = "#meetings", file type = .txt, .docx, or .pdf
3. Action: Run agent with the file content as input
4. Output: Post summary in #meeting-summaries channel
Option C: Scheduled Trigger
1. Trigger: — Every Friday at 5 PM
2. Action: Agent reads all meeting transcripts from the past week
3. Output: Weekly meeting digest sent to the team
Step 5: Build a Multi-Agent Team
Now let's level up. Instead of one agent, we'll build a team of three agents that work together:
1. — Parses raw transcripts, identifies speakers, timestamps, and segments
2. — Produces the structured summary (your existing agent)
3. — Creates calendar invites for action item deadlines, sends reminder emails, and updates the project management tool
Set Up the Team
1. Go to
2. Name: "Meeting Intelligence"
3. Add agents:
- Add your existing Meeting Summarizer
- Create new Transcript Analyzer (use the "Data Parser" template as a starting point)
- Create new Follow-Up Agent (use the "Task Manager" template)
Design the Workflow
In the visual workflow builder, connect the agents:
```
[Input: Meeting Transcript]
↓
[Transcript Analyzer] → Structured data (speakers, segments, timestamps)
↓
[Meeting Summarizer] → Summary + action items + decisions
↓
[Follow-Up Agent] → Calendar invites + task creation + email reminders
↓
[Output: Summary document + updated project board + calendar events]
```
Configure Team Settings
- Sequential (each agent waits for the previous one to complete)
- If any agent fails, notify the team lead and save partial results
- All agents share a common memory space for this task
- 15 minutes for the entire team
Test the Team
Run the team with a real meeting transcript and verify:
1. The Transcript Analyzer correctly identifies speakers and segments
2. The Summarizer produces a high-quality summary using the structured data
3. The Follow-Up Agent creates calendar invites and tasks correctly
4. The entire pipeline completes within the timeout
Step 6: Explore the Marketplace
The AgentWork Club marketplace is where you discover agents built by the community and share your own.
Browse and Install
1. Go to from the dashboard
2. Browse by category: Productivity, Customer Service, Research, Development, Marketing
3. Filter by rating (4+ stars), downloads (1,000+), and compatibility (your connected LLM)
4. Click "Install" on any agent to add it to your workspace
5. Installed agents appear in your Agents list with a "Marketplace" badge
Publish Your Agent
Once you have a well-tested agent, share it with the community:
1. Go to your agent →
2. Add a compelling title and description
3. Choose categories and tags
4. Set pricing:
- Anyone can use it (great for building reputation)
- Basic version free, advanced features paid
- Set your monthly price (AgentWork Club takes 30% commission)
5. Submit for review (quality check takes 1-2 business days)
6. Once approved, your agent appears in the marketplace
- Include a clear demo video showing the agent in action
- Write detailed documentation with examples
- Respond to reviews and questions promptly
- Update regularly based on user feedback
- Start free to build reputation, then introduce paid tiers
Step 7: Deploy to Production
Production Checklist
Before deploying agents that interact with real users or real data:
- [ ] Tested with 50+ diverse inputs
- [ ] Edge cases handled (empty inputs, very long inputs, special characters)
- [ ] Error handling tested (tool failures, API timeouts, rate limits)
- [ ] Guardrails configured (runtime limits, cost caps, output filtering)
- [ ] Monitoring and alerts set up
- [ ] Rollback plan documented
- [ ] Security review completed (data access, permissions, PII handling)
Deployment Options
1. Run from the dashboard or API when needed
2. Run on a cron schedule (e.g., daily at 9 AM)
3. Trigger from external systems via HTTP POST
4. Trigger from Slack, email, or other connected tools
API Deployment
For developers integrating agents into applications:
```bash
# Install the SDK
npm install @agentwork/sdk
# Or for Python
pip install agentwork-sdk
```
```typescript
import { AgentWork } from '@agentwork/sdk';
const client = new AgentWork({ apiKey: process.env.AWC_API_KEY });
// Run your meeting summarizer
const result = await client.agents.run('meeting-summarizer-v1', {
input: transcriptText,
options: {
timeout: 300, // 5 minutes
saveOutput: true,
notifyOnComplete: 'user@example.com',
},
});
console.log(result.summary);
console.log(result.actionItems);
```
Monitoring
Once deployed, monitor your agents:
- Real-time agent status, success rates, and costs
- Detailed execution logs for every agent run
- Configure alerts for failures, cost overruns, and performance degradation
- Weekly performance summaries emailed to your team
Step 8: Advanced Features
Custom Tools
Build custom tools that your agents can use:
```typescript
const customTool = {
name: 'company_database',
description: 'Query the company database for customer information',
parameters: {
query: { type: 'string', description: 'SQL query to execute' },
limit: { type: 'number', description: 'Max rows to return' },
},
execute: async (params) => {
// Your custom logic here
const results = await db.query(params.query, { limit: params.limit });
return results;
},
};
// Attach to your agent
await client.agents.addTool('my-agent-id', customTool);
```
Webhooks and Integrations
Connect agents to your existing tools:
- Agents can read messages, post responses, and react to events
- Agents can create issues, review PRs, and manage repositories
- Agents can read/write docs, sheets, and calendar events
- Agents can query and update CRM records
- Connect any system via HTTP webhooks
Collaboration Spaces
Create shared spaces where multiple users and agents collaborate:
1. Go to
2. Name it and add members (humans and agents)
3. Agents in the space can be @mentioned to contribute
4. All activity is logged and searchable
Getting Help
Documentation
- docs.agentworkclub.com — comprehensive guides and API reference
- community-contributed agent configurations for common use cases
- YouTube channel with step-by-step walkthroughs
Community
- discord.agentworkclub.com — 5,000+ members, real-time help
- community.agentworkclub.com — searchable Q&A
- github.com/agentworkclub — open-source SDK and tools
Support
- Community support (Discord, forum)
- Email support (48-hour response time)
- Priority email support (24-hour response time)
- Dedicated support engineer + Slack channel
Your Next Steps
Now that you have built your first agent and team, here are paths to explore:
1. Identify a repetitive task in your daily work and build an agent for it
2. Introduce yourself in Discord, share what you're building, get feedback
3. Install 2-3 agents from the marketplace and see how others approach problems
4. Share your Meeting Summarizer (or whatever you built) and get community feedback
5. Combine agents from the marketplace with your own custom agents
AgentWork Club is more than a platform — it's a community of people building the future of AI agent collaboration. Welcome to the club.