Quick Start

First pipeline in
under 5 minutes.

You'll install n00dles, define three agents, chain them into a pipeline, and run it. No YAML, no classes, no configuration files.

Install n00dles

Run this in your terminal. Requires Python 3.10+.

bash

pip install n00dles

Set your API key

n00dles reads credentials from environment variables. Add this to your shell profile or .env file.

bash

# Anthropic (recommended)
export ANTHROPIC_API_KEY="sk-ant-api03-..."

# Or OpenAI
export OPENAI_API_KEY="sk-proj-..."

ℹ

Don't have a key? Get one at console.anthropic.com — Haiku is fast and cheap for testing.

Define your first agent

Create pipeline.py. An agent is just a decorated function — the docstring becomes the prompt.

python — pipeline.py

from n00dles import agent, pipeline, run

# One decorator. That's the whole API.
@agent(model="claude-haiku-4-5")
def researcher(topic: str) -> str:
    """Research the given topic. Return 3 key facts as bullet points."""

Chain agents into a pipeline

Add two more agents and wire them with >>. n00dles passes the output of each agent as input to the next.

python — pipeline.py (continued)

@agent(model="claude-sonnet-4-6")
def writer(research: str) -> str:
    """Write a 200-word article based on the research."""

@agent(model="claude-haiku-4-5")
def editor(draft: str) -> str:
    """Polish the draft. Fix any awkward phrasing."""

# >> chains output → input sequentially
content_pipeline = pipeline(
    researcher >> writer >> editor,
    retry=3,
    timeout=60
)

Run it

Add run() and execute your file. n00dles will call each agent in sequence, retrying on failures.

python — pipeline.py (final)

# One line to run the whole pipeline
result = run(content_pipeline, topic="The future of multi-agent AI")
print(result.output)
print(f"Ran in {result.duration_ms}ms — {result.total_tokens} tokens")

bash

python pipeline.py

✓

You should see your article in the terminal within a few seconds. Each agent's trace is printed automatically.

Deploy to production

When you're ready to ship, noodles deploy wraps your pipeline as a serverless function and gives you a live endpoint.

bash

noodles deploy pipeline.py --name content-pipeline
# → Deployed to https://run.n00dles.io/your-org/content-pipeline
# → POST requests trigger your pipeline
# → Logs and traces in your dashboard

ℹ

Need a noodles account? Sign up free — no credit card required.

First pipeline inunder 5 minutes.

First pipeline in
under 5 minutes.