Quick Start

Deploy your first autonomous AI agent in under 60 seconds. No infrastructure setup, no DevOps required.

Prerequisites

All you need is:

• A GitHub or GitLab account
• Your agent code in a repository
• 60 seconds of your time

Deploy Your First Agent

# Or use the CLI
npm install -g swiftclaw-cli
swiftclaw login

What Just Happened?

Behind the scenes, SwiftClaw:

• Provisioned Infrastructure: Dedicated compute resources in a secure environment
• Configured Auto-Scaling: Your agent scales automatically based on load
• Set Up Monitoring: Real-time dashboards showing performance and errors
• Enabled Multi-Model Support: Switch AI models without redeployment
• Configured Persistent Memory: Context persists across sessions

All of this happens automatically. No configuration required.

Testing Your Agent

Once deployed, you get:

Production API Endpoint

curl https://your-agent.swiftclaw.app/api/chat \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Hello, agent!",
    "session_id": "user-123"
  }'

Real-Time Logs

View logs in the dashboard or via CLI:

swiftclaw logs your-agent --follow

Performance Metrics

Monitor your agent's performance:

• Response times (p50, p95, p99)
• Success/error rates
• Cost per request
• Active sessions

Example Agent Code

# agent.py
from swiftclaw import Agent, Message

agent = Agent(
    name="my-first-agent",
    model="gpt-4",
    memory="persistent"
)

@agent.on_message
async def handle_message(message: Message):
    # Your agent logic here
    response = await agent.generate(
        prompt=message.content,
        context=message.session.memory
    )
    
    return response

if __name__ == "__main__":
    agent.run()

// agent.ts
import { Agent, Message } from 'swiftclaw';

const agent = new Agent({
  name: 'my-first-agent',
  model: 'gpt-4',
  memory: 'persistent'
});

agent.onMessage(async (message: Message) => {
  // Your agent logic here
  const response = await agent.generate({
    prompt: message.content,
    context: message.session.memory
  });
  
  return response;
});

agent.run();

Next Steps

Now that your agent is deployed:

Add Multi-Channel Support

Deploy your agent to messaging platforms:

swiftclaw channels add slack --agent your-agent
swiftclaw channels add discord --agent your-agent

Configure Advanced Memory

Set up custom memory patterns:

swiftclaw memory configure your-agent \
  --short-term-ttl 3600 \
  --long-term-retention 30d \
  --vector-search enabled

Deploy Multiple Agents

Run multiple agents for different use cases:

swiftclaw agents create customer-support --model claude
swiftclaw agents create data-analyst --model gpt-4
swiftclaw agents create code-reviewer --model gemini

Set Up Monitoring Alerts

Get notified of issues:

swiftclaw alerts create \
  --agent your-agent \
  --metric error-rate \
  --threshold 5% \
  --channel email

Common Questions

Can I deploy multiple agents? Yes. Deploy unlimited agents on paid plans. Each agent runs in isolation with dedicated resources.

What if I need to update my agent? Push to your repository. SwiftClaw automatically redeploys with zero downtime.

How does scaling work? Completely automatic. SwiftClaw scales your agent based on load without manual intervention.

Can I use my own AI models? Yes. SwiftClaw supports custom model endpoints and self-hosted models.

Getting Help

• Documentation: docs.swiftclaw.com
• Community: Discord
• Support: support@swiftclaw.com