neonloops 0.1.1

Python SDK for Neonloops — run AI workflows via API

neonloops

Python SDK for Neonloops — run AI workflows via API.

Installation

pip install neonloops

Quick Start

Async (recommended)

import asyncio
from neonloops import Runner, RunInput

runner = Runner(
    api_key="nl_sk_...",
    # base_url="https://neonloops.com",  # optional, defaults to https://neonloops.com
)

async def main():
    result = await runner.run(
        workflow_id="wf_abc123",
        input=[RunInput(role="user", content="Hello, run my workflow!")],
    )
    print(result.output)
    print(result.metadata.tokens)

asyncio.run(main())

Sync

from neonloops import Runner, RunInput

runner = Runner(
    api_key="nl_sk_...",
    # base_url="https://neonloops.com",  # optional, defaults to https://neonloops.com
)

result = runner.run_sync(
    workflow_id="wf_abc123",
    input=[RunInput(role="user", content="Hello!")],
)
print(result.output)

Configuration

Parameter	Type	Default	Description
api_key	str	—	Required. Your Neonloops API key (nl_sk_...)
base_url	str	https://neonloops.com	Base URL of your Neonloops instance
project_id	str	None	Default project ID to scope requests
timeout	float	120.0	Request timeout in seconds
max_retries	int	2	Max retries for 429/5xx errors

Error Handling

from neonloops import Runner, NeonloopsApiError, NeonloopsTimeoutError

try:
    result = await runner.run(
        workflow_id="wf_abc123",
        input=[RunInput(role="user", content="Hello")],
    )
except NeonloopsApiError as e:
    print(f"API error {e.status}: {e}")
except NeonloopsTimeoutError as e:
    print(f"Timeout: {e}")

License

MIT