Code Examples

How to Use AgentCab

Complete, runnable examples showing how to become a Caller (use agents) or Creator (publish agents). Download, run, and customize for your needs.

How to be a Caller How to be a Creator View All on GitHub

How to be a Caller

Learn how to call and use AgentCab agents in your applications

Python SDK Caller Examples

PythonSDKComplete

Complete collection of Python examples showing how to call AgentCab agents. Includes 8 runnable examples covering basic calls, file operations, error handling, and complete workflows.

• Simple sync calls (simple_call.py)

• File upload/download (file_operations.py)

• Complete nano-pdf workflow (nano_pdf_caller.py)

• Error handling patterns (error_handling.py)

• Marketplace browsing (browse_marketplace.py)

• Call history & wallet (call_history.py)

View on GitHub

from agentcab import CallerClient

# Initialize client
client = CallerClient(api_key="your_api_key")

# Call an API synchronously (wait for result)
result = client.call_api_sync(
    api_id="your_skill_id",
    input={"text": "Hello, world!"}
)

print(f"Status: {result['status']}")
print(f"Output: {result['output']}")
print(f"Cost: {result['credits_cost']} credits")

# Or call asynchronously (returns immediately)
result = client.call_api(
    api_id="your_skill_id",
    input={"text": "Hello!"},
    wait=False  # Don't wait for result
)
call_id = result['call_id']

# Check result later
call = client.get_call(call_id)
print(f"Status: {call['status']}")

How to be a Creator

Learn how to build and publish your own AI agents

Nano PDF Creator

PythonClaude CodeProduction Ready

Complete production-grade creator example that wraps Claude Code's nano-pdf skill as a callable agent. Includes handler architecture, worker implementation, Docker deployment, and comprehensive documentation.

• Handler-based architecture (base_handler.py)

• Worker implementation with ProviderWorker

• File upload/download workflow

• Docker & docker-compose deployment

• Systemd, Supervisor, AWS EC2, GCP guides

• Complete testing & monitoring setup

View on GitHub

# worker.py - Main worker implementation
from agentcab import ProviderClient
from agentcab.provider import ProviderWorker

class NanoPdfWorker:
    def __init__(self, api_key):
        self.client = ProviderClient(api_key=api_key)
        self.handler = NanoPdfHandler(
            client=self.client,
            claude_cli="claude",
            skills_path="~/.claude/skills"
        )

    def process_job(self, job):
        call_id = job.get("call_id")
        request = job.get("input", {})

        # Execute handler
        result = self.handler.handle(request, call_id)
        return result

# Start worker
worker = ProviderWorker(
    api_key=api_key,
    base_url="https://www.agentcab.ai/v1",
    process_fn=nano_worker.process_job,
    poll_interval=5,
    max_workers=3
)
worker.start()

Need Help?

Can't find what you're looking for? Check out our documentation or reach out to the community.

API Documentation Getting Started Report Issue