Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.hud.ai/llms.txt

Use this file to discover all available pages before exploring further.

Automations are pre-configured scenario runs. You pair a scenario with a model and optionally pre-fill arguments, then run it whenever you need. Each run produces a traced result. A scenario qualifies as an automation when all its arguments are user-provided — no platform-injected args like trace_id or hud_api_key.

​
How to Use

​
From the Platform

  1. Go to the Agents page
  2. Click New Agent and select Automation
  3. Select an environment → scenario → model
  4. Pre-fill any arguments you want baked in — leave required args blank for a partial automation that prompts at runtime
  5. Your automation appears under Your Automations on the agents page
  6. Click it to view config, run it, or see run history

​
From the Environments Page

  1. Open any environment and load a scenario
  2. The “Make Automation” button lets you jump to the agents page to create one

​
Building a Scenario for Automations

Any scenario without trace_id/hud_api_key args is automatically classified as an automation: 
from hud import Environment

env = Environment(name="my-env")

@env.scenario("generate_report")
async def generate_report(
    topic: str,
    format: str = "markdown",
    max_length: int = 1000,
) -> Any:
    """Generate a report on the given topic."""
    prompt = f"Write a {format} report about {topic} (max {max_length} words)."
    response = yield prompt
    yield 1.0 if response else 0.0
When you create an automation from this scenario, you can pre-fill topic and format so each run just executes with those defaults. Arguments you don’t configure at creation time won’t appear in the saved config — they’ll be asked at runtime instead.

​
See Also